The very low budget DIY synergy horn build - Page 2

This is the first attempt I've seen at using a dome tweeter. I wonder if you might get a better result with a Vifa ring radiator along with an active crossover. What would also be interesting would be running just one mid converted into closed back, vs the same unit as you have now. You can then compare the sim vs measured in REW and see how they compare. You can then see how well Hornresp predicted your result.

VERY cool! I would love to try a build like this.

Very interesting thread, I like that your aiming to 100Hz, makes them usable with subs or towers. I'd love to see them stay passive, but it could become a pain!

Have you investigated what effect different shapes will do to response and directivity?

Subscribed.

Yeah, it can, use the 2x4 with the "4-way Advanced" crossover plugin. You'll need a 2x4 for each speaker, though.

You could possibly move the ports to the far sides of the horn. 1 port per woofer, and 2 woofers, right? If so, consider this: top face, port left, and bottom face, port right. This will give a bit of symmetry.

John

Thanks ... I got together with Gainphile and Antripodean (who has a Yorkville Unity) and measured them outdoors for directivity. The results are here:

http://redspade-audio.blogspot.com.a...surements.html

What you see with conical Synergy horns is that they have constant directivity down to a point defined by their coverage angle and mouth size, which is predicted by a formula that I believe Bill put into his spreadsheet. Having a big mouth and a wide angle extend a constant beamwidth down lower, but this must be balanced against size and extension. A narrower coverage horn tends to extend the bottom end down further, or if you want you can add more drivers to have a wide horn that also goes low for a given size. This is why I use 6 mids in a 2 way, it allows me to get down low with an 80 x 60 whilst using small mids that extend up past 1k. It's all about choosing your compromises!

The other thing you notice is that sometimes the pattern narrows at a certain point before pattern control is lost and then broadens at low frequencies. This effect was described by Keele in his paper on the CE horn. This is why you see a second wider flare section on Synergy horns - the idea is to avoid that narrowing. If you download his paper on the CE horn, which is basically an exponential throat joined to a conical horn with the latter 1/3 of the horn at 2x the flare angle, you will see it's a bit more dramatic than the second flare on Synergy horns. This is something to experiment with, although it takes quite some time to keep building multiple versions of this thing!

You may notice the Yorkville doesn't do so well in this regard and they did not bother to pay much attention to this aspect. It has a bit more top end as they used a BMS compression driver but being a fairly small axi symmetric conical horn, the directivity isn't as good as the others, it's about the same size as a 15" driver 0.4m. S1 which is the next one is 0.5m wide (60x60) and S2 is 90 x 40 and 1m wide, so extends beamwidth down to 280 Hz.

When I heard the original Unity kit, it had about $700 worth of passive crossover in it! Some decent quality parts in it, but for that you can buy digital xo and amps!

Looking at that mid response, if doing it active you could just add a couple of notch filters and then xo around 300 / 900. That then leads to a compression driver and again if you are doing it active this is easier. Celestion CDX1 is a good one to consider. At one GTG we measured and compared it to a B&C DE250 and it measured very similar and the sound also appeared quite close. I'd guess it would be fine with a crossover that low for home use.

Regarding your woofers, the holes are placed in corners as you probably realise to avoid interference. Having two holes allows them to be smaller and avoid different path differences from the cone to the port. However, in this case I'd overhang the baffle so the woofer has just one port in the corner, and I would change the shape to a slot if size becomes an issue. If using two woofers only, then you want to mirror the other woofer and have it over the same side of the horn. The idea is to have vibrations from drivers working against each other. If these are cheap buyout woofers and you have enough, perhaps just use 4? You then overhang the baffles and then brace the edges that extend out against each other.

CDX1 ... hmmm you got me there! It was a bolt on ferrite magnet version. Looking it up, would have to be this one:

http://www.parts-express.com/pe/show...umber=294-2124


There is another:
http://www.parts-express.com/pe/show...number=299-283

8 of them for $40 + shipping ... you could pick them up with a CD and the shipping might look better.

I haven't tried a 3 way horn, nor have I measured putting ports in the middle, so it's hard to comment beyond a very rough guess. I wouldn't try it unless you want to find out why not to do it the wrong way!

You can see even in some Danley designs where they show a 2 and 3 way version on the same horn, woofer ports have an impact. That becomes even more so when you break the rules!

Pattern control is probably not at stake here, that is a function of the design of the horn itself primarily.

$700 for both.


Don't overlook the fact that a passive crossover only works for one attempt, while a digital active crossover can be recycled many times over. Judging by your work so far, you'll get your money back and come out ahead. Spend a little on a few amp channels and a digital active xo and you can quickly put together different designs, optimise them better over time and not have to buy parts for each version of the crossover. You can always build a passive xo once you reach a point you are happy to stick with something for a while.

I've been involved in events where we have set up a range of active speakers in a day, swapped out some drivers, mixed and matched, measured a bunch and listened. We rigged up crossovers on the spot in minutes.

One example here:

http://redspade-audio.blogspot.com.a...-shootout.html
and another
http://redspade-audio.blogspot.com.a...ide-gtg-2.html

If this is an experiment then consider doing one as you have planned, and another with the ports in the corners.

I'm working on a 3way too, and am up against a similar conundrum...do I make big round ports for the woofers or use oval holes radially aligned? Which does the least damage?

Interesting point. The Yorkville has no corners, so they use slots for the mids as well. The response is reasonably smooth. In the case of woofers I would simply use a slot, although it is curious that DSL don't always do this, probably where they have determined they aren't necessary.

The biggest benefit you'll appreciate from a minidsp is delay. You can quickly time align your drivers and get a flat phase over a wide range of coverage, instead of having to design a ladder delay network.

The other thing you'll find out is that you can add eq boost, not just cuts that notch filters yield. This will allow you to "prop up" the frequencies around your cuts to keep things nice and flat. Plus, room eq wizard has an "easy" button- just choose your target and let it do the heavy lifting.

I know $100 bucks a pop isn't chump change, but I don't think you'll regret it once you've used it.

FWIW,
John

Lookin' good. How are you going to get a crossover point over 1k? With the measurements posted, it looks like your upper F3 is around 850hz or so. I'm sure I'm missing something here.

I'll add this other observation, and anyone can chime in here with their thoughts: smoothing the throat of the horn is KEY! This is what I've gathered from reading lots of diyaudio posts by Earl Geddes and Tom Danley. From what I can tell, any sharp transition points (like right angle corners) in the first inch or two are major diffraction generators, which can generate high order modes which contribute to "horn honk" and "bite." Here's a recent relevant thread:

http://www.diyaudio.com/forums/multi...mpressive.html

Turning a square hole into a round one that then transitions into a square horn is the job at hand here. I really liked using drywall mud for my horn, as its non toxic, easily sandable (you can do it with your finger), easily fixable if you goof, dries/layers quickly (especially if you use a heat gun), and can be reinforced with drywall sealer. I can't speak to the longevity of this method, however, as I only just built my horn.

Please note that I'm not a speaker designer or in any way an expert on horns. The above just represents what I've boiled down from the experts over the past several months of reading.

John

Really nice work on the throat, and it shows in the graphs. What kind of wood rasp did you use? I used ovoid and rat tail files. Can you post a pic of the rasp?

John

If you make the square of the throat smaller, so the edges overlap the circle, then you will make things easier, with less filling and a little more filing.

What are your port offsets?

You can import a hornresp sim into REW and see how your accuracy compares. You export, delete two colums in Excel (2nd and 3rd from memory), save it then import into REW.

Looks like you will struggle to get the crossover higher than about 850 Hz, unless you can get the mids to extend higher before working on the crossover.

This is the kind of transition you might want to aim for:



You can see the 18mm mounting plate where there is a radius transition that takes into account the compression driver exit angle. The idea is to make it smooth and not increase the area too rapidly. The fact that you are going from a circle to a square tends to make the increase in area more rapid. I don't think I have any photos or images that show how I actually did it, but you can imagine the walls set in a bit more.

I'd say that is really close.

That's a pretty good correlation. You can export into Akabak and then also import that into REW as well. I found Akabak did better with the same details above the upper rolloff point.

My conclusion so far with the upper end of the mid is that you just have to allow for a bit more than you think you need.

Me, too. Looks like you can drop in a notch filter at around 3.6k or so w/ a Q of maybe 1.5-2 and kill that hump, taking the low pass of your crossover from good to great.

John