DIY Synergy horn

For anyone not familiar with the Synergy:
http://redspade-audio.blogspot.com/2...ergy-horn.html

This project has been on hold for a long time while I attempted to learn Akabak. Then I realised I could sim this in Hornresp. I found a driver that seems like it will work - Pyle PDMR5. My sims:



I had to guess some parameters, but after doing some quick measurements I think I've come up with a better estimate. Turns out that BL and mms are pretty important - high BL and low mms brings the top end up. In that sim above, the response is falling due to those being less than ideal. After adjusting them (my second guess), the top end comes up a bit.

From the sims, I've figured out a few things. (Well, I take that with a grain of salt). It appears that more drivers (6 instead of 4) allows wider dispersion and more extension. The Unity which was my starting point gets down to 300 Hz. Making the horn bigger can extend it lower - I'd like more like 250 Hz to make it easier to match up with the 18" woofers. The efficiency drops down a bit - again 6 drivers help here. Also, this thing wants to be 60 degrees! If I make it wider, I start losing the bottom end. Again, more drivers can help here. 6 per side in a 60 x 90 could work well.

I'm currently building my first prototype, which is a 60 x 60 conical with a mouth that is about 0.5 x 0.5m. It came as a shock that the mitres weren't as easy as I thought to figure out!

Woofers that will go below them:
http://redspade-audio.blogspot.com/2...e-woofers.html


So the basic concept I have in mind is a Synergy horn which is built into a very large bass trap. It will look like a studio style soffit arrangement, with the 18" woofers below and some (future) tapped horn subs below the floor. My first step is to get a prototype version going that works well. I'm planning to try different ports and different experiements to learn about how this thing works. If it all goes well, then next step is a more serious construction CNC cut. The trap combo brings big speakers out of the room, otherwise I wouldn't be happy with the chunkiness sitting out there.

Building the horn was more difficult than I had expected! On the weekend, I worked with dad on it, and I didn't have the plan ready, so we had to figure the cuts out first. Dad tackled it using the manual method ie cutting a mitre on a sample, finding it isn't right, then adjusting til it works. I tackled it on 3D cad by intersecting the 4 walls then subtracting to let the computer work out the angles. Both methods came up with very contradictory results! The manual method said about 33 degrees for a 60 degree conical. CAD said 14 degrees! We ditched the CAD because the physical one we saw right in front of us working.

Later after cutting 3 sides, I realised why the results were different. In CAD, half the edges were mitred. In dad's manual method, all of them had the same mitre. So against expectation, the CAD resulted in something easier to build! I don't know if this will make sense to anyone except those who have built similar, always hard to visualise from words.

So I tried a couple of "thought experiments" in hornresp. What happens if I use 6 driver? Or non tapered ports? or 90 vs 60 degrees? Or the offset too large?

Here is 4 vs 6 drivers:


The others are in here:
http://redspade-audio.blogspot.com/2...p-lessons.html

One more sim, this one is really interesting - Misco JC5 vs Pyle PDMR5:



Both have the same horn, the Pyle has 6 drivers and the parameters have been tweaked. The Misco is a bit better when compared fairly, but it's also much more expensive. 12 Pyle drivers costs less than 8 Misco. I have heard the original Unity and the midrange caught my attention.

What you can see in that plot is the result of fiddling with the port area, length and the volume of air around the cone. Increasing the amount of air under the cone reduces out of band output. Making the port length a bit longer (in this case 9mm) helps. There dip then a peak after the top end rolls off. In most designs I come up with the peak is quite high. Adjusting those 3 parameters (port length, area and front chamber volume) can lower the peak and bring it closer to the dip. They tend to cancel each other out and combine into a smoother roll off that reduces the out of band output.

Why does that matter? The beauty of the bandpass is that it filters acoustically. That means the filter reduces harmonic distortion, which is then trapped in the chamber. The higher distortion products which do more harm are filtered to an even greater extent, and research has shown that they count the most. So the measured distortion is really really low. I think this is part of the sound of the Unity midrange. Distortion probably biased towards the more benign lower order products.

The question I have to answer is whether the tweaking in the sims produces a real result, or do I have to tweak the physical version. It takes a lot more time to try different things in the actual prototype! Seconds turn into hours. I hope to have my first working prototype this weekend, not sure if I'll pull that off.

Back to that plot. 1k > 2nd order distortion at 2k. The Misco version is 3 db down @ 2k - I'm talking about the acoustic filtering of distortion products here. Pyle version 17 db down, so about 15 db less. 3rd order - the Misco is about 12 db down while the Pyle about 25 db. By 5th order the Misco is 18 db down and the Pyle has dropped well below the floor. So whatever the distortion might have been based on a horn with no front chamber, it's now much lower again. This is one thing I like about the Synergy horn - taking some otherwise mediocre drivers and getting them to do something spectacular. But it's a risky project that could also end up a huge flop!

Noah, this will have a compression driver at the apex, otherwise I would be using a single mid driver in a more conventional horn. I didn't mention the CD because there isn't much to that part of it. Using multiple smaller drivers is only needed to keep the entrance holes small to avoid messing up the HF response. Most Danley Synergy horns load bandpass woofers into the horn. In my case I'm using just one but the front port will be right below the mouth of the mid horn. Try fitting a single 18" woofer into a Synergy!

The holes are 86mm offset (1/4 WL @ 1k). Crossover will be digital with DCX. In terms of the physical aspect, there isn't much to the CD. It can't move! The opening is a given and obviously the waveguide/horn profile will affect the directivity, which is one of the main things I'm interested in. The prototype will be conical but in the final I'd like to make it oblate spheroid - little difference will be seen by the mids as only the reflected part will be seen by the mids. In other words, there isn't a lot of juggling with this part of the design, hence I'm focusing more on all the options around the mids.

Thanks for the suggestions regarding the mitre.

Thanks for your comments Mark. One thing that would be nifty is to be able to change the offset without having to work out the area each time. CD is B&C DE250. I haven't tried it with a 1k xo, but I'd like to get it as high as I could. Crossover will be digital active (DCX) so that should help.

It seems there is a sweet spot in the placement of the holes. The offset seems to affect to a great extent how the mids couple to the horn. Move them to the apex, not good. Move them further down, also not good! Danley has mentioned the offset being a key aspect of him coming up with the whole concept.

Speaking of bandpass, WinISD doesn't help much there! The large predicted peak didn't show up.

Wi,
The problem I see is losing the top end, although some 1.4" drivers seem to be able to get up there. What seems ideal (and expensive) is a driver like the BMS 2" coax which can apparently get down to 500 Hz. A pair costs more than I'm spending in total on drivers, and I'm a bit surprised that Danley hasn't used it.

My idea with OS is to transition to the pyramid conical by the time it gets to the holes. Planning on getting the final version done on CNC. The idea is all in my head at the moment, haven't quite worked it out.

I already have 4 x DE250, 2 for this, 2 for surrounds. The 2nd pair were cheap second hand, too good to pass up.


Not sure if I understand your question. If the sims prove accurate then I'll do 6 mids and probably one slot each for the drivers on top and bottom, those slots being in the corners, then the other 2 in the sides done the normal way, 2 holes right in the corners.

I'd like to do 60 x 90 but the sims suggest I lose the bottom end, but 60 x 60 seems to work well. That's what I'm trying first.

I'm thinking CNC with laminated sheets, a circle to square transition with the OS radius. I may be relying in part on the CNC guy, or I may have to figure out how to model it. I don't plan on getting into formulas, just looking for the smoothest transition I can get. I might be using some filler as well, I don't think the CNC will do it perfectly, it may be stepped. My test version I'm building now is conical.