Yep!
I agree it needs a lot of work. I'm nowhere near making a mold.
Should the fleece be saturated completely in one shot? I hit it with resin on the throat plug side but it only fully saturated in a few spots. The back almost looks untouched. I figured I'd hit it on the back side as well. Would it help to use better resin? I've been using Bondo brand. It's pretty thick.
I made another one to hopefully address the issues I was having a couple weeks ago.......namely the 14khz off axis notch and the widening pattern around 4khz. The change in dispersion hasn't improved. I can't get the material to stretch enough to smooth out the wrinkles without it changing angle where my throat plug ends. The plug is about 3" deep. I bondo'd and sanded the transition smooth (it wasn't bad to begin with). The off axis notch at 14khz was due to the gap between the cd and my wg throat from the gasket. Plumber's putty took care of that.
I can't quite figure out what's going on from 3khz-9k. It looks pretty chaotic to me, with the off-axis curves criss-crossing all the way out to 30°. Is this some type of diffraction? I printed out the os curve that I used to generate the profile in cad, and my wg follows it quite well......it just has a little bit more curve. Not much though. Maybe it's nothing to be too concerned about?
The verticals look strange too between 1khz and 2khz. Waistbanding before the loss of control? It's 13.5" wide before the mouth rolls back so it should have control to 1khz.
The first set of data is horizontal, then oblique (rotated about 45°), and vertical.
I can't quite figure out what's going on from 3khz-9k. It looks pretty chaotic to me, with the off-axis curves criss-crossing all the way out to 30°. Is this some type of diffraction? I printed out the os curve that I used to generate the profile in cad, and my wg follows it quite well......it just has a little bit more curve. Not much though. Maybe it's nothing to be too concerned about?
The verticals look strange too between 1khz and 2khz. Waistbanding before the loss of control? It's 13.5" wide before the mouth rolls back so it should have control to 1khz.
The first set of data is horizontal, then oblique (rotated about 45°), and vertical.
Attachments
Yeah, the vertical is something you give up going with an elliptical wg. I was hoping that it was large enough to overcome this. Still not too bad.......I'd be looking to cross around 900hz with these.
I never considered asymmetry as far as the criss-crossing goes. I've taken care to make these as symmetrical as possible, though my construction method certainly isn't perfect. That said I don't see how the opposite side is going to effect the wave moving down the waveguide like that. I would think of that as diffraction and these are quite smooth. Maybe I'm not getting what you're saying
. Wouldn't be the first time!
I'm looking at a polar fr plot of Geddes' old Summa. The quality of the pic is poor, but they appear to show the same behavior past the on-axis dip, just not to the extent that I have.
I never considered asymmetry as far as the criss-crossing goes. I've taken care to make these as symmetrical as possible, though my construction method certainly isn't perfect. That said I don't see how the opposite side is going to effect the wave moving down the waveguide like that. I would think of that as diffraction and these are quite smooth. Maybe I'm not getting what you're saying
. Wouldn't be the first time!I'm looking at a polar fr plot of Geddes' old Summa. The quality of the pic is poor, but they appear to show the same behavior past the on-axis dip, just not to the extent that I have.
I would say that you are seeing some mouth diffraction - you have to expect that unless your radii gets much larger. And there is no guarantee that an elliptical profile like this can ever be "perfect" - its just not an exact coordinate system so odd things are bound to happen. Overall, it looks quite good - it may not ever get any better.
Yeah, the vertical is something you give up going with an elliptical wg. I was hoping that it was large enough to overcome this. Still not too bad.......I'd be looking to cross around 900hz with these.
I never considered asymmetry as far as the criss-crossing goes. I've taken care to make these as symmetrical as possible, though my construction method certainly isn't perfect. That said I don't see how the opposite side is going to effect the wave moving down the waveguide like that. I would think of that as diffraction and these are quite smooth. Maybe I'm not getting what you're saying
I'm looking at a polar fr plot of Geddes' old Summa. The quality of the pic is poor, but they appear to show the same behavior past the on-axis dip, just not to the extent that I have.
The data in polarmap is quite recent and quite high resolution. You should look at that for comparison. Send me the data and I can post it online. A polarmap is always more enlightening to me - it takes some time, but once you "get-it" polarmaps just shows things the way they are.
Yeah, I think I'll go ahead and make a mold from this. I've been chasing my tail the last couple days trying to make them better but the changes have been almost insignificant, once I had the throat right. I like looking at your Polar Map data.......I've been looking at mine in ARTA (called a Filled Contour Plot there), but I wouldn't see the criss-crossing in the contour plot. I like FR plots for details like that. Maybe I'm not skilled enough at reading the contour plots yet.
Have you thought about adding side injection ports to turn it into a synergy? You could fiberglass some plywood driver mounts for the mids and woofers and maybe get something like the Redspade fiberglass horn. You will have to grow the design a bit to support the CD down to below the mid XO circa 700-900Hz.
http://www.redspade.com.au/audio/PSE-144.php
http://www.redspade.com.au/audio/PSE-144.php
Last edited:
I think I mentioned it in the first post, but my intention here was to build these "small" waveguides to test my build method before I dive into the Synergies. Once I started these smaller waveguides i decided to push it as far as I could, and to make good use of my TD15s. I also have some cardioid midwoofer cabs in the works that will go with these waveguides to maintain directivity from 300hz+.
I've been modeling a big 3' wide 2-way synergy for a couple years now. Real close to what Paul has done but mine will be more like 60x40. I've got all the drivers. Next year!
I've been modeling a big 3' wide 2-way synergy for a couple years now. Real close to what Paul has done but mine will be more like 60x40. I've got all the drivers. Next year!
Earl, here's the data imported to Holm. It was taken in REW so there's no time sync. I didn't include all the data points I had because I had to go through every text file and remove the header to get it into Holm. Lazy on Saturday morning I guess so I just included the angles that you specified in your measurement paper. Your paper also said you gate the data in your sw so this is the raw response.
so I just included the angles that you specified in your measurement paper. Your paper also said you gate the data in your sw so this is the raw response.
Hi Nate
Unfortunately that is not enough data. It stops at 80 and needs to go to 100, 120, 150- and then 180. It may seem like these rearward points are not very important, but they are, because the model is of a source in free space. I have tried in the past to "simulate" the rearward data so that I only needed to take forward data, but then the approach that I use doesn't work so well. I hope getting that rear data isn't a big problem.
Unfortunately that is not enough data. It stops at 80 and needs to go to 100, 120, 150- and then 180. It may seem like these rearward points are not very important, but they are, because the model is of a source in free space. I have tried in the past to "simulate" the rearward data so that I only needed to take forward data, but then the approach that I use doesn't work so well. I hope getting that rear data isn't a big problem.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.