Thought I'd see what people think about a little two way I've made before spending money on crossover components. Nothing too special, a Dayton DSA 135 woofer and SB26 STCN tweeter. I assumed there would be a dispersion mismatch and there is. There is also a notch in the DI at 350hz that I'm unsure of it's cause. Speaker was measured outdoors on a 10ft ladder. I have plenty of experience with active filtering but wanted to try passive for various reasons, one being to verify my measurement process.
Reverse null
Reverse null
Most amps should be able to cope with the way you have shorted the amp via the woofer circuit at high frequencies, but not everyone would feel comfortable doing that without introducing a little resistance into the mix.
As mentioned, add a resistor in series with that bypass cap on the 2.5mH coil.
I would consider allowing the woofer to go slightly higher to fill the mild depression at 2k.
I would also see how it looks using gated measurements.
Looks like full BSC. Are these going on stands, away from walls?
I would consider allowing the woofer to go slightly higher to fill the mild depression at 2k.
I would also see how it looks using gated measurements.
Looks like full BSC. Are these going on stands, away from walls?
Both the HP and LP have declining behavior in impedance at higher frequencies. Its significant enough to cause audible distortion and be on the edge of partial oscillation instability with some amps.
The SB26STCN is a good little tweeter that doesn't need complex filtering in most design situations. It does have higher H2 than most, but higher order HD is relatively low and doesn't show tendency to climb at higher output levels.
My only assumption of the DI dip at 350 hz is destructive interference between woofer and port output (If so equipped).
The SB26STCN is a good little tweeter that doesn't need complex filtering in most design situations. It does have higher H2 than most, but higher order HD is relatively low and doesn't show tendency to climb at higher output levels.
My only assumption of the DI dip at 350 hz is destructive interference between woofer and port output (If so equipped).
I would consider adding a resistor in series with the 4.7uf cap on the tweeter. That's likely what is causing the impedance issue profiguy just mentioned.
Is there any particular reason you want to cross so low? It would be easier on the tweeter to cross higher.
Is there any particular reason you want to cross so low? It would be easier on the tweeter to cross higher.
Thanks guys. Resistor additions noted, seem simple enough to implement I added one to the tweeter here. The amp I use with these is a crown xli800.
I don't tend to consider baffle step compensation, I simply filter the speaker in a manner that would create something flat in an anechoic setting. Any boundary issues I resolve with EQ as best as possible.
With a nearfield use case and fairly low SPL, I don't see it as much of a problem. I didn't really register any distortion when doing some sweeps in room with active filtering. With that said, I have raised in in this iteration below.
I will admit I'm a bit confused on the phase graph here. It doesn't quite behave as expected. There may be an issue with my data. When measuring the phase is kind of a mess so I cleaned it up with the "Estimate IR Delay" action in REW. I discovered this reading another thread but I'm not entirely sure this is an appropriate action.
I don't tend to consider baffle step compensation, I simply filter the speaker in a manner that would create something flat in an anechoic setting. Any boundary issues I resolve with EQ as best as possible.
With a nearfield use case and fairly low SPL, I don't see it as much of a problem. I didn't really register any distortion when doing some sweeps in room with active filtering. With that said, I have raised in in this iteration below.
I will admit I'm a bit confused on the phase graph here. It doesn't quite behave as expected. There may be an issue with my data. When measuring the phase is kind of a mess so I cleaned it up with the "Estimate IR Delay" action in REW. I discovered this reading another thread but I'm not entirely sure this is an appropriate action.
What is the reason for tweeter Z -18mm? It would mean tweeter is physically closer to listener.
If the baffle is tilted ot stepped, Z should be positive value.
And in the case where drivers ACs are aligned, it really makes sense to use LR2, here around 3kHz as starting point.
And I see 0 for Y of both drivers? How did you obtain the measurements? Common mic position or each driver separately in its height?
If the baffle is tilted ot stepped, Z should be positive value.
And in the case where drivers ACs are aligned, it really makes sense to use LR2, here around 3kHz as starting point.
And I see 0 for Y of both drivers? How did you obtain the measurements? Common mic position or each driver separately in its height?
That's the woofer with the offset. The tweeter IS closer to the listener on this speaker.
It is a flat baffle.
Single mic position for both drivers.
Common mic position, both driver on the same flat baffle, Z must be 0 for both drivers.
0-90 hor 10 degree increments, no vertical
FWIW I live pretty much right next to a highway.
But you were measuring with a reference timing channel? That should not have been necessary and even counterproductive. Be sure to keep the start of the time window in the impulse response constant and (of course) left of all your impulse peaks.
I think the xo looks fine, but I would make this suggestion. In your sim, try different values that are 10% higher and lower to see if any values are finicky. For example, the 1.2mH coil. See how far each direction you can go, before the response is affected significantly. Maybe a .8mH works just as well. Or maybe a 1.5mH does. Then see what is a commonly available value, or one that you already own.
I'm still curious what it looks like with gated measurements. Might not change much as you are outdoors, and up high, but my indoor measurements change a bit.
I'm still curious what it looks like with gated measurements. Might not change much as you are outdoors, and up high, but my indoor measurements change a bit.
Last edited: