So, I didn't want to hi-jack Bill's thread regarding his Cosyne horns. His spreadsheet, plus the numerous reviews, plus my fascination with Danley invented technologies have inspired me to build some of my own. I'm starting this thread with the intent of it being a build log. Now, with the two little girls my wife and are trying to adopt, the progress will be slower than I'd like I'm sure. So be patient if there are gaps. I'm sure along the way I'll have to back-track for mistakes, but that's part of the learning process. First my design. I'm building fronts for a home theater. My gracious wife has allowed them to be pretty much any shape and size as long as: "They aren't as big as those big subs you used to have when we got married. " Those were Adire 15" DVC woofers in 3'x3'x3' enclosures. Something tells me the horns are going to end up roughly 22.5"x22.5"x20", maybe a little larger...so I'll keep my end of the bargain. My room is longer than it is wide, and because of the projection screen the speakers end up being pretty close to the side walls. A 90x60 horn would have required a really steep toe-in to avoid wall reflections. I'm going with a 60x60 horn to avoid those reflections. I'm going to post my jigs and methods for folks to get ideas. I used to have my own speaker company, and I've built a bit of furniture/cabinetry. Don't take my word as gospel on any of these methods, they may or may not work for you. I do know that I've used them very successfully in the past. First, I made a small set of test cuts on scraps, to make sure that the angles I got off an online compound miter calculator were correct, and also to help me visually figure out how to lay the horn out for butt-joints which I'll be able to screw and glue for greater strength, and better alignment. I make a lot of drawings on pieces, and take a lot of notes right on the pieces. I learned long ago, that this does take time, but completely re-making a piece because you got confused takes longer. Next I made a template piece for what will become the "A" piece, or inside piece of the initial horn flare. The "B" piece will be much more complicated due to the fact that it will have "wings" on it to make room for the bass drivers. (More on that later.) I started by making a full size layout on the piece. I figured out the length by making a scale drawing in Visio, and I cut that on the table saw. The angle cuts were measured, not set up using a protractor. A fraction of a degree at the beginning of a ~18" line will yield unacceptable errors. One thing to keep in mind, sharpen your pencil often. Wide layout lines are of no use whatsoever, and can lead to gaps of 1/32" to 1/16" easily. I will keep 10-20 sharp pencils at my disposal and sharpen them all at once. Next I made a zero-clearence sled for the circular saw.
How the skill saw guide works. Note...make sure you can make a sacrificial cut into the surface you're working on. Make sure your work-pieces are not going to go flying off. I will use a combination of clamps and brad nails. I actually just nailed this to the pieces I was cutting, and I'll fill the holes later. The second picture shows the bevel sled I made, which is still under construction. The piece set on top is the master "A" piece I cut using the zero clearance guide. With a couple of modifications that will be used as the master pattern for the rest of the pieces. I'll be cutting those on the table saw using a special fence setup. That will have to wait though...since we have to take the girls to the park now.
Very cool! Sounds like what youre making will resemble an SH50? Have you thought of mounting them against the walls to make them effectively twice as wide? Danley suggested that in some forum post (sorry dont have the link).
Full three way, with most likely quad 8" woofers. Drivers I'm considering so far are: Dayton PA130 Mids, Dayton Polyimide Compression tweeter, and possibly a quad of the Dayton Classic 8" woofers. The price/performance of all those drivers are outstanding. The quad 8"s would have more surface area than a single 15", with a reasonable though not spectacular 4.5mm xmax. I just wish they made an 8 ohm version of the classic sub 8" with 8.5mm xmax and a copper cap on the pole piece. I'm trying to stick with all currently available drivers.
SH60, but with directivity only down to ~550-650 Hz. Keeps it much more livable. I've seen the wall mounting thing before but that might take them a bit wide, plus I'm not sure the wife unit would go for that. I'll be using quad 8" woofers, because I can't get a 12" to fit. (I don't think I can, still working that out.) I might actually be able to get dual 10" to fit...it's not the width that is the problem it's the distance from mid to secondary flare. Bill, I blame you for this hobby you know. First Liberty Audiosuite, then Praxis. (I never could swing an IMP in college.) Now you've gotten me into Synergy style horns? You are a co-dependent enabler. Scott
Okay now for some sims...been working a bunch in Horn Response by David McBean. Pretty sure I owe him a bottle of wine or 12. Fantastic software. Unbelievable value. ;-) So I'm using Dayton PA130 mids...which seamed to have a good mix of maximum output, parameters suitable for a midrange driver and cost. I've had great experiences with the Parts Express house brand drivers before from their reference series, to the classic series and everything in between. Also they aren't all that expensive and there's a good chance they'll still be available in a year or 4, which can't always be said for buyout drivers. (I have bought those before too...) My one concern is that they wouldn't go quite high enough to maintain linear phase, but we'll cross that bridge when we get there. (I'm not there yet.) One thing I've noticed in horn response simulating what seams like thousands of possible choices is that the volume between driver and horn, or throat chamber, and port area/length all set up a resonance which appears to have a right volume. I'm not sure if that's actually the case...but it sure looks like it. This is what the response looks like in my simulated horn with the "right" volume, too small and then too big.
Too small... Notice how the direction of the resonance changes. Too big and it goes up first...too small and it goes down first. Now in real life, I'm not sure how the various energy storage (or Q) parameters impact this, and I don't know how well David models them. I know that OD type horns are pushing the envelope so it will be interesting. I plan on making molded plugs to adjust volume using Bondo, or maybe a professional molding agent. ($$$) I measured the cone volume by taking an extra midrage and putting a layer of cling-wrap down. Then....I filled it with water. 100 cubic centimeter's exactly.
Oh, and the crappy Trendnet power line Ethernet adapter is broken. They have a poor trace that carries mains currents...it's fused on 2 of my 4 units. Sigh. The ~25cc's I need means I need to fill the cone area with bondo up to the level of the surround. The PA 130's have the classic voice coil lead in wires coming through the cone, which then get's insulated with a polymer. I'll have to account for that to make sure they don't smack into the bondo plug at high volumes. No wonder these things cost a fortune. ;-)
Hi Scott: Impressively smooth simulations. I've been doing a lot of sims also so I can appreciate that result. Let's compare notes. My target is a 90 x 40 horn for a corner location. It shows 5db passband ripple unless I leave off the final flares - then it smooths out nicely. One thing I'm wrestling with is do I need them with the horn pushed back tightly into a corner? Because of the corner, that narrowing near where the horn looses control shouldn't occur because the corner maintains control. So I think I will just try to smoothly blend the wings into the wall and apply roundovers to the top and bottom edges. There is a lot of anecdotal evidence on the web re' why the flares but I think I'm going to have to try to know for sure. I noticed that Wayne Parham doesn't use them on his corner horns... I'm simulating with HiVi BN4. 4" 3.3 mm xmax. otherwise similar to the Dayton. Can be placed closer to the apex but needs to cross at 200 Hz or so based on excursion. In terms of tailoring the acoustic low pass, I've seen a lot of sensitivity to the throat aperture size and port length and lower senstivity (or at least broad bands of insensitivity) to chamber volume and area. But I'm not looking at volumes that need phase plugs to be realized so maybe that is why. I'll take a look at that (since you are pointing the way); it might help. Thanks for the idea/picture on the sled for cutting the angles. You make it look so easy I'm starting to believe that I can do it! Thanks for sharing all this. And, of course, thanks also to Bill Jack
I'm glad this has been helpful to a few folks. I've learned a lot studying these and lurking. So here are my thoughts on the ripple, I've done a lot of research on this and yet to come up with a definitive answer. 1. See the attachments for a Keele like flare ~1/3 at double angle and a Visaton FRS7, which can be placed very close to the throat of the horn. Tons of ripple, even though in the Keele paper [SIZE=78%]http://www.xlrtechs.com/dbkeele.com/PDF/Keele%20(1975-05%20AES%20Preprint)%20-%20Whats%20So%20Sacred%20Exp%20Horns.pdf[/SIZE] he says the purpose is to flatten out the response over the coverage angle. Even Bill Waslo's simulations and build show a fair amount of ripple for his midrange drivers. 2. There is an entire thread on selecting a midrange driver on DIY Audio. Basically some of the horn gurus there use a modified variant of a rule of thumb that's pretty classic for front loaded midrange horns. For the Synergy tapped horns the derivation ends up being that you want 2XFS/QES to be ~600 or so. The PA130 only manages a 341, while the FRS7 yields 351. In my simulations drivers with a relatively high FS (low moving mass) and high BL do better. Variants of the Celestion and Misco closed back drivers that Thomas Danley is known to have used have this high BL and High FS (they also meet the criterion of the 2xFS/QES from what I've read) 3. That ripple looks an awful lot like diffraction ripple. And...I believe Horn Response models the conical horn as if it were a circular cone, not a squared off cone. I've measured what must be 30 or 40 different DIY and professional designs where a relatively thick flanged tweeter is not flush mounted. Something tells me that when measuring a slight bit off axis, or for a rectangular cone it won't be as bad, but I won't know for sure until I measure. 4. I purposely reduced the size and angle of the flare on my final expansions. That helped minimize ripple and increase the high frequency response. My thought process is I'm not trying to cover a few hundred seats at a stadium, I'll probably end up sitting in quite a bit less of the coverage area. I can tune for a sweet spot. 5. I looked at HiVi driver as well, but with it's higher mass, lower BL, stiffer suspension I couldn't get it to go as high as the PA130 Look at my last sims, that's PA130 vs. HIVI in my chosen expansion rate. Black is PA130, grey is HIVI 6. Last but not least, I believe I've figured out a way to be able to try different flares on my prototype I'm working on. I could be mistaken, so I reserve the right to change my tune at any time. ;-) Scott
One last thought. My "a-ha" moment for both Tapped Horns and Synergy horns came from realizing both designs work by pressurizing a space to launch a wavefront someplace other than the driver itself. In the case of a tapped horn it's constructive summation of a single driver over a given range. (Or multiple drivers over a range, just essentially in parallel in the same air space). For Synergy it's pressurizing a horn by using symmetrical entry points over the length. The entry points are where it's convenient for driver geometry and the summation for best phase response. This may not be accurate, but this is how I understand it. And now for geometry. I was able to lay out the "B" or outside piece of the first flare of my horn tonight. The 4 year old was very helpful, she sat on my outfeed table and handed me tools like it was surgery. She also tried to explain to me "The Ladybug and Penny" game. I still have no clue what that was, but it made her happy to explain it. It looks like 4 8" woofers will be the order of the day. I could lengthen the horn, have a wider flare and fit 4 10" woofers or 2x 12". 4 8's however have a surface area a bit larger than most 15's. The Dayton Classic 8" has reasonable throw. If someone can find me a sub $30 8", 8 ohm woofer with either 6-8mm of throw or a copper sleeve/cap, I'd be willing to try it out. I probably won't have much time for construction until Saturday. Tomorrow is ballet (for the 4 year old, not me) and Friday I'm solo Dad. I will spoil them rotten and let them fall asleep on the couch watching movies. It's just easier that way. ;-) Scott
One last thought for the night...the reason I made the comments about pressurizing space for the drivers: Both designs seem to work better with high motor strength drivers. The ability of the driver to push the air load in front of it seems to be critical to performance. With tapped horns it's not always a realizable design (too long, to much compression ratio at S2) but it does seem to be the case. Once you give up looking at frequency response, and off axis response of the driver and think of it as a piston pushing on an air mass, then, at least for me it all fell into place. Scott
Scott, FWIW the ripple on mine has never been as much as modeled. I think the losses and/or asymmetries smootb out the peaks. Have you looked at the Silver Flutes at Madisound?
Yeah, I noticed that, which again led me to believe that it's an artificially enhanced artifact of the modelling. I bought several of the 6.5" silver flutes (I think I still have 4) when I had my speaker company. Nice drivers and great values for sure. The 5.25" versions model very similarly to the PA130. The 8" would require a trim ring to clear the surround, which could be done. It's one to keep in mind for sure. Scott
Instead of a trim ring you could instead use a router and circle jig to cut a moat in the outside of the horn wall for the surround to clear. Helps keep the front chamber volume down, too.
I'd probably have do a combination of both. The 6.5" Silver Flute drivers have a surround that sticks up ~5mm over the frame face. It looks like the 8" is even more, possibly 8mm or so from the pictures. (It's tough to judge though.) With another 6mm of xmax, and probably 5-10mm of mechanical travel before suspension limits, that makes the MDF or plywood dangerously thin. The 8" Dayton Classic Sub driver (more expensive than I'm willing to use for this, since it would be 12 of them) has in upwards of 20mm mechanical throw in one direction. It's a beast of a driver, if it shares any of the clearances with the much less expensive Classic 8" woofer driver I'll have to employ both techniques. Scott
Scott: Thanks for the detailed responses. I need some time to absorb them and my day job is getting in the way. I will keep the conversation going on two points at least - the use of that mass load corner figure of merit for mids and whether your PA-130 can indeed go higher than my smaller BN4s, The smaller mids can be placed closer to the apex and have small differential path lengths from various areas on the cone to an off center throat port.... Jack
I hope I'm wrong, the figure of merit means there are relatively few very good drivers for the application, small enough to fit where you want them at the throat, with the right parameters. What do you have as a measurement system? Scott
I think you are correct about the difficulty of finding a good driver but I also think the problem is significantly eased when one uses (as Danley recommends) a driver suitable for a 1 Khz or lower crossover frequency. In all the simulations of candidate synergy mids I've been doing, the high frequency limit has been determined by the reflection from the throat. So I pay less attention to the metric and more to what a quick simulation tells me about whether it looks good enough - at the low end of the range as well as at the high end. The 1/4 lambda cancellation causes me to wonder where exactly the reflection is coming from. My reluctant conclusion is that its past the CD mounting plate and potentially deep inside the compression driver itself. In the 4550 (which I don't yet own), it looks like it is 1" or so deep into the driver. When I add that 1" to the L12 of the horn, I see my margin going away but not quite gone. That is where the advantage of a smaller diameter driver comes into play, as I'm sure you know. It's center can be placed closer to the throat and there is less additional potential HF rolloff introduced by an off-center throat hole when its necessary to drill the hole off center in order to get it close enoght to the throat. I'll have to mount the mid to a test hole at a diameter for which I have a plug cutter, make a measurement to determine the acoustic path length, and then adjust as needed and move on to the tuning of the mid chamber and port. I have a Dayton MIC with the dayton supplied calibration curve and am comfortable using REW and ARTA. I'm sure they will get a workout on this project. Re' tuning the chamber and port. In the Unity patent, Danley seems to say use the rolloff from the cancellation and the acoustic low pass as part of the crossover and the driver spacing to make the phases align. In the Synergy patent he recommends overlap between drivers as making it easy to get the correct phase characteristic. But how much overlap? This has a big bearing on how high the mid needs to go and therefore how close to the throat the port holes need to be. I don't know how to answer that question myself other than by trying to design the crossover and so I anticipate another round of cut and try if/when I get to that point. Don't get me wrong. I don't really know what I'm doing. I'm just an oldguy newbie with a couple of subs and an active SEOS design under his belt.I do have an EE degree and I understand the design process. I love the idea of getting my hands dirty again and working with components that I can actually see with the naked eye. I'm humbled by your and Bill's evident knowledge and experience in the field. I learn a lot from these build threads and its great to be able to talk to others about my new hobby. But I don't want to hijack your thread so I'll let you go back to the build. Jack
I'll be able to get some work done on it tomorrow. I had solo dad duty tonight while my wife was at a writers group meeting. Ahem. Girls night out is more like it. I won't get to measurements for another week or two. The concept of a 1/4 wavelength cancellation being a bigger limit is an interesting one. A quick numbers check would indicate that the cancellation would occur at somewhere between 1200 and 1600Hz depending on the acoustic termination of the compression driver. I'll have to look for a notch there not in the simulation. I can also offer to measure with and without the compression driver I'm planning on using (Dayton D250P) to watch the cancellation move. Tomorrow my plan is to finish the bevel sled, then work out the details on the "B" piece template. We'll see how far I actually get. Scott
Got pretty far today. First pic is the drawing of the "B" piece template. The way I did it the biggest side of it is the inside of the horn. Took me a bit to work that out. For the "A" pieces I used a template technique for the table saw. Creates way less mess than a router and is much easier to do. There are a couple of catches with the technique: 1. Make several trial cuts to get it lined up just right. If there is still a gap move the fence closer to the blade. Try not to cut the template since you'll want to use it for outlining pieces to cut later. (Possibly.) 2. Make sure you aren't pinching a piece against the fence. That's a recipe for kickback. And since most speaker-builders are guys. And the kickback would hit you with a rather small piece in the...um...guy parts....that's bad. 3. Go slow. This isn't like cutting regular boards, I wouldn't do this with hardwoods that might bind unless I had the splitter in place. I would stick with MDF and plywood.