[Saurav]

Now on to the crossovers

Drivers are AE OB12, Audax PR170M0, BMS 4538/18Sound XT1086.

[MJK]

Saurav,

Interesting design and driver combination. I will be interested in the crossover/EQ developments as you proceed. They came out very nice looking.

Martin

[Saurav]

Thanks. I built these with a jigsaw, and this is my first attempt at staining anything, so overall I'm satisfied with how they turned out. I didn't upload the photos which highlight the crookedness

I've been running crossover sims for months, now it's time to start building and measuring. I'm planning on a passive XO between the mid and tweeter, and that'll be driven by my SET. Active XO+EQ to the woofer, with an SS amp on that. So far the passive's been the bigger headache. The tweeter is about 4" deeper than the midrange, so I've been struggling with getting the phase to look halfway decent. I have a prototype that I'll start with and then see where I end up.

For the woofer the measured results on a prototype baffle matched your MathCAD sim very nicely. I have the shelving filter worked out, and I'll wait till I have final measurements on this baffle before I decide on whether I need the notches or not. I got an active XO PCB that will eventually have the whole thing, but I'll start with using that just for the shelving filter (and notches if needed), and use a Behringer for the 'straight' XO portion. That'll make it easier to play with different XO points and see what I like.

This is going to keep me busy for a while

Once I started thinking about it, I realized that the design goals for this project are kinda haphazard:

* I wanted to use that midrange because I had it in my previous speakers (the blonde baffles in the background) and like the way it sounds, and to keep the initial costs down

* I had a ribbon tweeter and never liked the integration, my midrange didn't go high enough and the tweeter didn't go low enough. So I wanted a new tweeter, wanted to try a compression driver, and decided to try for some level of 'constant directivity' in the front. So the baffle width was chosen to keep the midrange response dipole-ish up to a higher-than-usual frequency, and then the crossover to the tweeter is at roughly where the midrange's response starts to narrow. I eyeballed these with off-axis measurements, so it's largely handwaving, but that was the goal. On the low end, because of the narrow baffle, the midrange starts to roll off around 500Hz, which is where I want to cross to the woofer, so I'm hoping I can hit my target slope without too much electrical help.

* I had the woofers in sealed cabinets in the old speakers, and wanted to try dipole woofers. This is where there seem to be two distinct schools of thought, and I went in the low-Qts/high-Xmax direction. We'll see how it works out. The woofer is pretty flat up to ~700Hz, so an LR4 XO at 4-500Hz should work OK. Because of that high XO, I didn't try a U or H frame, since any meaningful depth would add a resonance at a lower frequency.

So... hopefully I'll have a somewhat constant dispersion in the front from the bass up to the treble, and the back wave will start out dipole and then fade out at the mid-tweeter XO. I'll put that in the room and see how it sounds.

[dyohn]

Nice horns, Saurav.

[Saurav]

Thanks

[Saurav]

Haven't got all the measurements off of my laptop yet, but have some data to share. I figured the midrange/woofer EQ would be of most interest here, since that's the part that's related to being on an open baffle.

My test baffle to measure the OB12:



Note the box of nails propping up the magnet

Predicted response using MJK's worksheet:



And my measurements. All measurements are smoothed to 1/16th octave to make the graphs easier to read.

Blue = Nearfield, mic roughly at the plane of the surround
Black = Calculated OB response (invert, delay by 1/2 baffle width, combine with nearfield)
Pink = Ungated ground plane measurement from ~1m
Purple = ~10ms gated ground plane measurement



This convinced me to stick with a flat baffle, and not try U or H frames, because I figured I'd get cavity resonance problems well below 500Hz.

I have an active crossover/EQ between the woofer and midrange, and the woofer EQ's schematic and transfer function are shown below. It's basically a shelving low-pass filter to counter the dipole rolloff, and a notch filter to handle the peak.




Responses. Blue is calculated - I took the calculated dipole response from before, and applied the EQ transfer function. Red is an ungated/farfield ground plane measurement of the woofer with the EQ applied. Both charts are smoothed to 1/6 octave, and levels adjusted to line them up.



Added to this is an LR4 crossover at 500Hz, but I don't have graphs of the final response.

Because of the high crossover frequency, the midrange only needed a notch, and the LR4. I have the woofer crossed to a sealed monopole subwoofer at 80Hz. This is what I'm listening to right now, and there's lots of tweaking to come. The passive crossover between the mid and tweeter needs work. I had trouble getting the phase to line up with the ~4" acoustic offset difference, so I want to try some more slopes/frequencies to see if I can do a better job than what I have now.

This is a 2m ungated measurement of the full speaker, smoothed to 1/16 octave, before I put in the midrange notch filter: