A project of Nathan or Abbey and subwoofers

[pjpoes]

The HP isn't truely 1st order, it has notch filters that steepen the response. The LP is 2nd order, but the Inductor is what Dr. Geddes specs. The dip isn't enough to be polarity, plus I tried switching it once, doesn't make it go away. I should check again and make sure its right, but the Tweeter is supposed to have inverse polarity to the woofer, and I believe that is the case, as its supposed to be.

I tried playing around with the LP filter myself once since I found the inductors to be out of spec slightly anyway. They were supposed to be 3mh and were emasuring 3.2mh. I unwound them until I measured 3mh and tried again. Then unwound them to 2.9mh, but didn't want to go lower than that, since once you undo it, its hard to get the winds back just as tight. I also tried a different Inductor I had that was 2.5mh, but that creates a problem at both 1.5khz and the 800hz area where there is a bit of a bump right now.

Dr. Geddes redesigned the crossover and I've asked about the schematic so I can try it, but he seems quite insistent I shouldn't bother. I'd guess that the upgrade makes a bigger difference than the same kind of money spent on other sorts of tweaks people do, but he seems to feel its a waste of money, or at least thats my impression. If he doesn't feel that way, Then PLEASE send me the schematic.

[jzagaja]

In my case (D-800Ti driver) - smaller the coil better the LP slope:

Imageshack - kkk.gif

[pjpoes]

Whats your polar response look like with each of those though? You can't just look at a dip at one point based on a simulation. Like I said, when I lowered the coil size, I found the response change to be problematic, and to add to that, I saw signs that the polar response was no longer smoothly transitioning into the horn, so I stopped messing with it. Basically, I'm going to trust that Dr. Geddes knew what he was doing when he designed it. If I have to live with a small dip of at worse -5db's at 1khz (that may or may not be a speaker response issue) to ensure a flat polar response, I will do so.

I again would caution you to comapring a simulated crossover response change to the inroom measurements. These aren't quasi-anechoic, they aren't the measurements you would use in crossover design, they are quickly taken in room measurements at a listening position 4 meters away.

[jzagaja]

I recall I had similar dip as yours but I use 2" coil woofer (for better midrange) and with 3rd order LP no problem at all. Simulation is only a simulation.

[pjpoes]

I understand your point, I'm just reluctant to make crossover changes based on in room measurements. Too many factors could be causing the problems. Again, you also have to consider the power response, the entire response across the polar area. If making changes fixes the response in one area but messes it up in others, you have to consider if that makes sense. Weight it against the effect of say a 3-5db dip at 1khz.

I looked at some other measurements I have from my last house, from each speaker, different angles, etc. I have no quasi-anechoic ones that are worth considering, the gated measurements are a mess, but basically this is what I found. The 1khz dip is always present, but it varies. The dip seems to be roughly 3-5db's relative to the output at say 500-800hz, but is sometimes as little as 1db relative to the tweeter. I don't believe that changing the crossover will help that unless I change to a 3rd order. The shallower slope of a smaller inductor will cause the 400-500hz range to go up. The Woofer uses a CR in parallel, so its possible that adjusting both the inductor value and resistor value would improve it slightly. If I went into all that, I'd just redo the entire crossover using the new design.

[jzagaja]

Always in my case with longer window such a dip was shallower. Funny thing is we speak about power response that we don't know I have made Excel spreadsheet with recursive PR calculation.

[DougSmith]

Originally Posted by pjpoes

Wish I knew, I've seen it in other manufactured horn speakers, maybe its a horn thing. Maybe its my room? Maybe I did something wrong when assembling the crossovers. I have no real idea. Dr. Geddes doesn't show that in his measurements, so I'm not really sure.

I see a small dip in the same region with My Abbeys (~1-1.5 kHz), and it is more pronounced at the prime listening position (exact center) than 2 ft to either side. So, when I average the measurements it is less obvious - but it is still there. I considered the new x-over (which Earl did send me the specs for) but I'm not sure if that would eliminate it. I ended up adding PEQ filter to bring it up a bit (which was easy to do with my receiver). It does make an audible difference because the ear is quite sensitive to small changes at that frequency. If I decide I don't like it in the long run I can always remove it with a few button presses.

Doug

[pjpoes]

Hey Doug, would you mind sending me the new crossover schematic/specs? I work for the University of illinois (illinois.edu) and my service name is "mpoes" that should be enough to get my email. If I still manage to get junk mail after that convolution I might cry.

jzagaja I do know my power response in that Dr. Geddes and myself have measured the polar response of the speaker. I'm a little confused by your comment, my specific point was that when I made certain adjustments I was finding it effected the response. I don't mean a simulation, I mean actual measurements.