What's the Easiest DIY Unity Horn Project?

[Patrick Bateman]

I've been really tempted to buy some drivers for another Unity horn project. But I already have a pile of drivers from *existing* Unity horn projects.

I wanted to get some feedback from other members who've built Unity horns.

Basically:

1) If you've built a Unity horn, are you happy with the way it turned out?
2) If not, what would you change?


Off the top of my head, I believe there are two or three easy ways to build a Unity horn, and a number of difficult ways.

Let's start with the difficult options:

Option one would be for me to buy some compression drivers, some midranges, and get my measurement rig working. The upside of this option is that I enjoy experimenting with Unity horns, as illustrated by dozens of Unity horn projects that I've messed around with over the past seven years or so. The downside is that 2/3rds of these didn't work so well, and I usually lost enthusiasm to finish them, particularly when it came to cosmetics.

Option two would be to build one of the Synergy horn designs from Paul Spencer. Although they're not much more difficult to build than the ones from Bill Waslo or Tom Danley, the Spencer designs require a midrange that isn't readily available, so that makes it tricky. On the plus side, I believe Spencer's choice of midrange has some advantages over the one used by Waslo. (Bill's midrange is two bucks; all drivers have limitations, and at this price point, they get kinda severe.)

Those are the difficult options. Now let's talk easier options.

An externally hosted image should be here but it was not working when we last tested it.

Option three would be to build Bill Waslo's Synergy horn.
Bill's Synergy horn is arguably the best documented one in the whole world. Bill's published the crossover, the dimensions, a spreadsheet that lets you build your own horn, lots of pics, etc.
My main concern about building this one is the midranges. I've uses the $2 buyout midranges in my own projects, and they sound 'grungier' than the Peerless or Tangband 2" woofers. I have a whole case of the Peerless woofers, and I've certainly wondered if it might make a suitable replacement for the two buck mids.

Option four would be a Lambda Unity horn.
If I did this, I'd use the Pyle sealed back midranges instead of the Misco mids, and I'd use the Parts Express clone of the B&C DE250 instead of the B&C DE25. The drivers are close enough that the modifications to the crossover would be fairly minor. This option is tempting for a couple reasons. First, I've used the Pyle midranges, and they're by far the best value of all the mids I've tried. In my opinion, they sound as good as the most expensive mids that I've tried in a Unity horn, and they sound better than the two buck specials from Parts Express. Basically you could spend a lot more on midranges and they won't sound much better.
I haven't heard the D250P, but if it's a close knock off of the DE250 than it should work just fine.
About the only downside to cloning the Lambda Unity is that it's a kinda dated design; Danley has evolved the design quite a bit in the 12 years since it went on sale.
Then again, I've heard the Lambda Unity myself, and it sounds better than any of the Unity horn experiments that I've done personally. I had a couple Unity horns that worked a little better in one aspect or another, but none of my Unity horns were better overall.

An externally hosted image should be here but it was not working when we last tested it.

Option five would be to build a front loaded horn for the Kef KHT1005 satellites. For the life of me I can't figure out why more people don't use these for horns, if there's a better 3" driver for under $50 I don't know what it is. You can get a set of five, plus a subwoofer and an amplifier delivered to your door for under $250. Thiele Small on these is similar to the Fostex and Faital 3" woofers. The Kef is a coaxial, so it would looks a bit like these Danley Synergy Horns once it's horn loaded:

Downside to the Kefs is that they're not going to get remotely as loud as options one through three.



If I *do* go with option one, I was thinking something like this:

Basically a couple of compression drivers in a Paraline and a pair of eights. Sort of similar to the Sound Physics Labs 'Runt':

 

[ra7]

Not much experience building these, but FWIW, I'm going to try build bwaslo's well documented coSynes this summer.

 

[bwaslo]

A problem with my current 'well documented' design is that the woofers are NLA. Someone cleaned them all out of both Parts Express's stock and Madisound's.

I'm working on designing with some replacements. Also replacing the 2"ers (since they're buyouts, too), should make PB happier! Though after testing the ones that looked best on HornResponse modeling, they don't appear to work as well (for smoothness and sensitivity) as the 2 buck specials. One note, I only use the 2" for a fill between the tweeter and woofer, it doesn't cover much range but allows me to get the thing working linear phase. The design can be done with just tweeter and woofer, but then the crossovers have to be a lot sharper, so no linear phase with that.

I still want the system to reach to below 100Hz - hence the woofer (200Hz is too high to jump to a subwoofer and I want to be able to place subwoofers for better room mode handling). The Pyle midranges are too big, I couldn't fit those on along with woofers unless the horn was a lot bigger (which wouldn't fly with my wife).

I have a better/easier way of cutting and assembling now, was in the process of writing it all up when the woofers went bye-bye. So I'm working to get some stock drivers working before coming out with new revised (and even more detailed) plans.

Thanks for the plug, Patrick!

 

[Pallas]

An externally hosted image should be here but it was not working when we last tested it.

Option five would be to build a front loaded horn for the Kef KHT1005 satellites. For the life of me I can't figure out why more people don't use these for horns, if there's a better 3" driver for under $50 I don't know what it is. You can get a set of five, plus a subwoofer and an amplifier delivered to your door for under $250. Thiele Small on these is similar to the Fostex and Faital 3" woofers. The Kef is a coaxial, so it would looks a bit like these Danley Synergy Horns once it's horn loaded:

Downside to the Kefs is that they're not going to get remotely as loud as options one through three.

I'm not sure the KEFs have the motor strength to take advantage of horn loading. Even the Q100 driver, which has a magnet bigger than some subwoofers - I imagine KEF originally intended to use neo, only to change plans when neo prices skyrocketed - has barely half the motor (BL) of the BMS concentric Danley uses in the SM60.

That said, if you do choose this route, use the Q100 driver or the KHT3005SE driver. The newer Uni-Q's with the radial phase plug are just much much better performers than the older Uni-Qs. Also, that HTS1005.2 has a teeny tiny tweeter that doesn't really do dynamics.

 

[Patrick Bateman]

A problem with my current 'well documented' design is that the woofers are NLA. Someone cleaned them all out of both Parts Express's stock and Madisound's.

I'm working on designing with some replacements. Also replacing the 2"ers (since they're buyouts, too), should make PB happier! Though after testing the ones that looked best on HornResponse modeling, they don't appear to work as well (for smoothness and sensitivity) as the 2 buck specials. One note, I only use the 2" for a fill between the tweeter and woofer, it doesn't cover much range but allows me to get the thing working linear phase. The design can be done with just tweeter and woofer, but then the crossovers have to be a lot sharper, so no linear phase with that.

I still want the system to reach to below 100Hz - hence the woofer (200Hz is too high to jump to a subwoofer and I want to be able to place subwoofers for better room mode handling). The Pyle midranges are too big, I couldn't fit those on along with woofers unless the horn was a lot bigger (which wouldn't fly with my wife).

I have a better/easier way of cutting and assembling now, was in the process of writing it all up when the woofers went bye-bye. So I'm working to get some stock drivers working before coming out with new revised (and even more detailed) plans.

Thanks for the plug, Patrick!

That's a good point about the buyout drivers. When I used them in my projects, I was running them about 2-3 octaves. (About 300hz to 1500hz.) And I wasn't using a highpass, which would exacerbate distortion.

It sounds like you're only using them for about an octave or so in your design.

What do you think about the Celestion compression driver? The Dayton D250P looks good, but it kinda bugs me that it's such a blatant rip off of the B&C. If the Celestion works well, I can't see a reason not to use it. (IIRC, power handling on the Celestion is lower, but that's of little importance for a home audio situation, where we'll likely give the compression driver little power.)

 

[tinitus]

how do you control the drivers 'overlap', to avoid phase 'suckout' or peaks ?

 

[bwaslo]

It sounds like you're only using them for about an octave or so in your design.

That's about right. Really just a 'filler' driver almost, something that can be placed close enough to the throat to reach easily past the crossover frequency. I imagine it would be terrible if driven full range!

What do you think about the Celestion compression driver? The Dayton D250P looks good, but it kinda bugs me that it's such a blatant rip off of the B&C. If the Celestion works well, I can't see a reason not to use it. (IIRC, power handling on the Celestion is lower, but that's of little importance for a home audio situation, where we'll likely give the compression driver little power.)

Well, I really like the sound of the CDXi-1445 Celestion driver (extremely natural, does its job without calling attention to itself. Though part of that is the Synergy "sounds like one driver" effect). The 1445 also has a relatively wide throat angle, which helps maintain the beam width at the high end, and it also extends to a little over 20kHz without a sharp drop off. But unfortunately it doesn't go very low - 2kHz is about it, and it appears consistency isn't great. One of the last ones I got has trouble reaching even that. I also tried its big brother, the CDXi-1745, which surprisingly didn't seem to go much, if any, lower! It was a little more sensitive, though.

The 2kHz low end and the difficulty pushing the mids much above 3kHz (and that seems to vary with angle) has me now working with the large DE250 type mylar drivers. I'm using a Denovo DNA-350, since that's what I have available. (I'm not bothered by the DE250 imitations - there are no patent numbers listed in the DE250, and none of the imitators are claiming theirs to be a B&C brand device). I wish these had a wider throat angle and went a little higher (if only for prettier graphs!) but I do know they are good-sounding devices so this isn't a big compromise. I do wish they weren't so heavy, though! All this ferrite hanging on the glue-assembled wood horn has me worried that it wouldn't likely survive much of a drop.

 

[bwaslo]

how do you control the drivers 'overlap', to avoid phase 'suckout' or peaks ?

By controlling phase in the crossover, like always. But in a synergy horn, you don't have to worry about combining them in-phase to keep a lobe facing forward. 90 degrees (or whatever gives you flat response) is fine, since the waveguide controls the lobe always in the forward direction. One of the numerous things that manages to inherently go right in a synergy horn.

 

[Patrick Bateman]

At this point, I'm leaning towards a variation on the VTC Paraline box, but for home use. Here's why I think it would be compelling:

1) In a Unity or Synergy horn, I find that the most difficult part to get 'right' is the midrange. Basically it's hard to get the mids to play up to the xover point of the compression driver. About a week ago it occurred to me that the dual compression drivers in a Paraline would facilitate a lower xover point. After reviewng the DSP settings from VTC, my 'hunch' appears to be correct. The VTC box with the single compression driver has an xover point of about 950hz, but the one with the dual compression drivers has an xover point that's almost an octave lower! (The 'real' xover point is a bit nebulous, because there's a ton of EQ in the DSP settings.)

I think this is very exciting news, because it dramatically simplifies Unity horns. We can *finally* have a two way Unity horn that goes down to 100hz. This makes e-v-e-r-y-t-h-i-n-g so much easier:

a) Instead of a humongous and expensive three way crossover, we have a relatively simple and easy to manage two way crossover, using DSP of course
b) Instead of dedicating the next six months of my life to tweaking inductors and capacitors we just buy a minidsp, measure with the mic, then set the xover and EQ accordingly
c) We can experiment with various xover slopes, for instance we can try transient perfect xover topologies. While you *can* do that with a passive xover, it gets expensive in a hurry due to the number of components involved and the bizarre response shape of compression drivers, not to mention their crazy impedance curves

Probably the most compelling reason to do a 'VTC' style box is that there are numerous drivers that will work with an xover point of 600 or 700hz, but there are very few that will work with an xover point of 1500hz. Paul Spencer had to have his Unity horn midranges special ordered from Celestion, and I had mine special ordered from Misco.

What I'm picturing would look like this:

Dual Dayton D250Ps in a Paraline, with an xover point around 700hz or so

An externally hosted image should be here but it was not working when we last tested it.

Woofers are Silver Flute 4ohm eights

Crossover is MiniDSP

Cost for each channel is $273. ($99 for minidsp, $100 for dual Daytons, $74 for dual Silver Flutes.)

 

[Melo theory]

Sorry to hijack the thread but Mr. Bateman's PM box is full.....

Dear Patrick,

I have noticed from some of your posts that you are into car audio. I am desperately trying to find a DVD head unit that encodes Dolby digital, DTS and pro logic II.

I was going to do a MS matrix wiring to use with a center channel but pro logic II seems to be the best solution.

Please PM me!

Thanx Patrick. Now, I have to go....."I have to return some video tapes"

 

[Patrick Bateman]

I am not aware of any.
My car does this in the stock system and it definitely sounds very nice. This is probably the best stock system I've ever heard. (It's not perfect, there's no bass, but it gets the midrange right and that's important.)

You might be able to find the head unit on ebay, it's the Lexicon head unit from a Hyundai Genesis.

If you're in the market for a new car, it's a steal, I think I paid something like $31K for mine Fit and finish of a Lexus or an Infiniti for the price of Toyota Camry.

 

[Melo theory]

Oh yeah, I do like that car. My buddy's Acura TL has Dolby Digital audio and it was amazing. I just bought a civic SI.

I don't know why multi-channel car audio fell by the wayside....
What better place to have a center channel? Everybody is off axis!
Arrrgggg! I'm going to pull my hair out!

 

[natehansen66]

I think this is very exciting news, because it dramatically simplifies Unity horns. We can *finally* have a two way Unity horn that goes down to 100hz. This makes e-v-e-r-y-t-h-i-n-g so much easier:

Having built a pair of Paralines, and having lived with them for a few months now I have to question your logic a bit. Building a quality set (or quad in your case!) of Paralines is a good bit of work if they're intended for hi fi. Obviously by now we know that a comp driver on a Paraline will not measure anywhere near flat. On mine there are some anomalies that can't be eq'd out.....though I admit that's probably more so due to my construction. The way the mids enter those VTC boxes looks pretty questionable as well when you consider the comp driver's response. You also have to build the horn shell like you would with a Synergy. How about the vertical? Would you keep the 10 deg or so of the Paraline, or would you "shade" the top and bottom Paraline sections for wider dispersion?

I don't see anything at all easier with a Paraline Synergy, just a different set of trade-offs. Mid selection I suppose but if you build it to play to 100hz you're just introducing worse trade-offs IMO as far as the mid entrance ports go.

If I could have a re-do, I would have built "regular" Synergies from the start. Would have saved a lot of other headaches. That said, I'm not knocking the Paraline. I've got mine to where they're sounding pretty good, and I've learned a lot in this process about horns and audio in general. That in itself has been worth it.

There are several proven Synergy designs floating around the web now, and with bwaslo's cut list spreadsheet things are even better. It's really too bad about those Celestions. If they made them available to the public I bet they could move a few through PE.

Has JLH made any progress on his Paralines? I'd like to see what he comes up with.......

 

[Patrick Bateman]

I've ordered some drivers, but I'll need to bust out the CAD program before I make a final decision on whether it'll use a Paraline.

I *definitely* want to use two compression drivers per side. If I could justify the cost, I would use four. With four compression drivers, we could definitely get the xover point down to 500hz. I think with two we should be able to get it down to 700hz.

There are a couple ways to use two compression drivers:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

option one: two compression drivers in a Paraline, a la the VTC EL210 and the Yorkville PSA1.

OR

option two: two compression drivers in a Unity/Synergy horn, basically by arraying two horns next to each other so that they function as a single unit

If I'm not mistaken, the horn in option two would have to be fairly small in order to get the compression drivers to sum constructively. For instance, if the center to center spacing exceeds 13.5" you're going to see comb filtering above 1khz. This is dramatically complicated by the fact that the devices have directivity. Basically I believe that there's going to be a juggling act; if the center to center spacing is too big, they won't sum constructively at the low end of the compression driver's output. If they're too tight, they'll interfere with each other in the compression driver's midband.

 

[fuji6]

Hi Patrick,
When you mentioned 4 CD's per-side I instantly though of an idea.
You could build another "Stargate" device WITHOUT the top plate. Then mount it upside down on top of the existing stargate device (using the existing top). Then turn it sideways and add whatever horn to it.

 

[fuji6]

Hi Patrick,
When you mentioned 4 CD's per-side I instantly though of an idea.
You could build another "Stargate" device WITHOUT the top plate. Then mount it upside down on top of the existing stargate device (using the existing top). Then turn it sideways and add whatever horn to it.

 

[Patrick Bateman]

Hmmm I'll have to try and visualize that!

 

[Patrick Bateman]

Here's a quick mock up of what the double Unity might look like. (Visualize the box flipped 90 degrees; my image host has a maximum width of 800 pixels.)

Basically two 40x40 degree Unity horns splayed together.
Since the device is designed to be used in unison, I'm having them share a single cabinet.
So basically the device has about 80 degrees of coverage horizontally, and about 40 degrees vertically.
Vertical coverage will collapse fairly early, because the box is only 30cm tall. So basically below 1133hz, the vertical pattern will broaden. This probably isn't the end of the world, since reflections off the floor and ceiling aren't as destructive as horizontal reflections.

One big question I have is whether the gap between the horns will cause major issues.

 

[fuji6]

Hmmm I'll have to try and visualize that!

Think of it as building a stargate, then building another one on top of it.
So the layers you would have are.
Bottom plate (entrances holes for the first set of CD's)
3 plates with the guts of the stargate.
Top plate (acts as the top of the upside down stargate as well).
3 plates with guts of the second stargate.
"Bottom plate" (entrance holes for the second pair of CD's)

I have no idea if it would work, but it was just an idea.

 

[Patrick Bateman]

Having built a pair of Paralines, and having lived with them for a few months now I have to question your logic a bit. Building a quality set (or quad in your case!) of Paralines is a good bit of work if they're intended for hi fi. Obviously by now we know that a comp driver on a Paraline will not measure anywhere near flat. On mine there are some anomalies that can't be eq'd out.....though I admit that's probably more so due to my construction. The way the mids enter those VTC boxes looks pretty questionable as well when you consider the comp driver's response. You also have to build the horn shell like you would with a Synergy. How about the vertical? Would you keep the 10 deg or so of the Paraline, or would you "shade" the top and bottom Paraline sections for wider dispersion?

I don't see anything at all easier with a Paraline Synergy, just a different set of trade-offs. Mid selection I suppose but if you build it to play to 100hz you're just introducing worse trade-offs IMO as far as the mid entrance ports go.

If I could have a re-do, I would have built "regular" Synergies from the start. Would have saved a lot of other headaches. That said, I'm not knocking the Paraline. I've got mine to where they're sounding pretty good, and I've learned a lot in this process about horns and audio in general. That in itself has been worth it.

There are several proven Synergy designs floating around the web now, and with bwaslo's cut list spreadsheet things are even better. It's really too bad about those Celestions. If they made them available to the public I bet they could move a few through PE.

Has JLH made any progress on his Paralines? I'd like to see what he comes up with.......

At this point, we have a couple people who've built Paralines telling me "don't build a Paraline." (That would be you and Art Welter.)

Here's some food for thought:

In January I attended the CES, and had an opportunity to listen to Don Keele's CBT array. I didn't like the sound. It sounded similar to the B&O Beolab, the one with the SAW acoustic lens. But the B&O is dynamic; the CBT was not.

My theory is that dynamic peaks in the CBT are basically 'rounded off' because the sound is radiation from a very large surface.

An externally hosted image should be here but it was not working when we last tested it.

For instance, if a 3" loudspeaker radiated an impulse into a room, the sound will reflect off of the walls, the ceiling, the floor. The sound from the CBT radiates from a much larger surface, and due to that, the troughs in the impulse are 'filled in' and the peaks are truncated.

Bottom line - the sound of the CBT was 'mushy' to my ears. (Just a subjective opinion of course, and I know that someone that bought a set of Gedlee speakers wound up selling them for CBTs.)

OK, now let's look at the Paraline.

An externally hosted image should be here but it was not working when we last tested it.

In the Paraline, we *also* have the potential for this problem. Because we have high frequencies radiating from a strip that measures about 13" in height.

Now, clearly that's not as astigmatic as a Keele CBT. But it's not ideal either.

And I'm finally 'grokking' that this may be the reason that:
1) Danley doesn't sell line arrays
2) Danley sells Genesis horns, but his praise for them seems a bit qualified
3) The SH50 is the Danley flagship

Now at this point, someone will likely say "just throw in the towel already"

But I still think there's hope for the Paraline. First off, I'm not convinced that the 'wiggles' in the high frequency response are offensive. My Paralines sound pretty nice; if I had to complain about anything it would be the sound of the midranges. (Not a big fan of the two buck mids, though Bill Waslo made a good point that they shouldn't be used for such a wide bandwidth as I do.)

In my next post, I'll add some info on how we might improve on the Paraline.