Coherent summing issues
I'm not sure what drivers you guys are using but some JBL drivers are wired so that positive voltage applied to the
black terminal gives positive pressure. Here are notes on JBL drivers that will help you identify how to wire them:
Also, remember that when the diaphragm in a compression driver moves away from the magnet, this causes
negative pressure at the throat. Some manufacturers label the red terminal positive for positive pressure, others for "forward" movement, i.e. away from the magnet, which causes negative pressure at the throat.
The confusion lies in the fact that the magnet and diaphragm are mounted "backwards", and the sound passes through the center of the magnet, which forms the throat of the driver. This is just something you have to keep track of on an individual basis, since some manufacturers do it one way, others do it the opposite, and some manufacturers do it differently from one model to the next.
The notch shown on the measurements made by Skywave Rider suggest the compression driver is probably wired so that positive voltage causes positive diaphragm motion, not positive pressure. Could be the woofer is wired "backwards" but I'll bet it's the compression driver. One or the other is backwards, anyway. So when reversing the connections, you have actually phased them properly.
There is another possibility, and that is you may have crossovers that modify phase in such a way to cause destructive interference. The classic case is adjacent second-order networks, one moves phase forward 90 degrees and the other moves phase back 90 degrees so at frequencies very near the crossover point, there is nearly full cancellation. Most people reverse connect one of the drivers to counter this situation.
The thing is, the whole idea of the system you're doing is to make off-axis response smooth in addition to having good response on-axis. So you have a higher standard to achieve than just getting summing right on-axis. Because of this, you want to try and make sure you cross your t's and dot your i's where coherency is concerned.
No loudspeaker system with more than one driver on a baffle can be coherent at all locations and at all frequencies. Somewhere, there is going to be destructive interference. But the idea is to place your nulls outside the coverage angle. A clever designer makes it impossible for nulls to form within the wall angle of his horn, or at least uses them to abbreviate the cutoff at the edge of the pattern. This is a good place to put the link to the post about vertical null angles:
That's why I like using asymmetrical horn flares in a system like this. They allow the relatively wide 90 degree horizontal angle to match the collapsing directivity of the midwoofer. And the narrow vertical angle matches the null angle from vertical spacing of the drivers. If you space everything right, the nulls form just outside the vertical coverage angle of the tweeter horn. In the crossover region, the nulls set the edge of the vertical pattern and at higher frequencies, the tweeter horn's wall angle maintains the relatively tight vertical control. Axisymmetrical horns won't do this, instead, the vertical pattern will widen back up above the crossover band.
So the next thing to do is learn where the acoustic centers are in the Z axis, the acoustic distance. This is what sets the vertical centerline, which biases the arc of coherent summing in the vertical plane. The nulls that form above and below this centerline are found at angles determined by the distance between drivers and the frequencies where they are both used. As we saw earlier in this thread, it's common to space the drivers so the arc is about 40-50 degrees wide in the crossover region, usually around 1kHz - 1.5kHz or so.
Setting the centerline for the vertical pattern is a function of driver position on the baffle and crossover slopes. The acoustic position of the compression driver is set back because of horn length, but the woofer is also set back a few inches from the baffle because of the shape of the cone. There are differences in the electrical and mechanical characteristics and also in the acoustic loading. The woofer is a direct radiator with higher inductance, the tweeter is horn loaded and has less inductance.
Knowing the acoustic position is important and difficult, but fortunately, if you crossover at 1kHz or so, wavelength is over a foot long. That means you have a range of about six inches where interaction between drivers causes positive summing and about six inches where it causes cancellation. Actually, I should say six inches where summing "tends towards" positive or negative reinforcement; The fact is, at this frequency range you won't see the notch from negative reinforcement until you're within about a two inches forward or back of the position that causes path length to be 180 degrees out of phase. But this, alone, will tell you a lot.
If you have a good measurement system that allows impulse testing, super. You have it easy, life is good. You can measure the drivers and see the flight time delay for each one. That will tell you everything you need to know and crossover design becomes a piece of cake. Well, maybe not completely effortless because you still have some work finding the best slopes to use to get summing right off-axis as well as straight on. But being able to see what you're working with makes things a LOT easier.
Even if you just have an RTA, you still have some visibility. You'll have to measure outdoors to make sure what you're seeing isn't a reflection. If you wire up direct and measure straight on axis, then reverse polarity to one driver and see a notch, then you know the drivers are aligned. This also is true if a symmetrical crossover is used. If you don't notice a clear notch when reverse connecting the drivers, then you can move a driver forward or backward a few inches to find the notch. Start first by setting the notch straight on axis, even if it means moving the tweeter forward or back from being flush with the baffle.
This is the
Altec alignment method, and it works pretty well. You can't know if you're in phase or some multiple of 360 degrees out of phase, but since we're working with a foot long wavelength, you can rule out multi-cycle shifts by physical alignment. You may (will) have some offset due to electro-mechanico-acoustic differences, but they won't create a foot long offset. So as long as you have the physical alignment close in a system like this, you have enough to learn where the center of your arc is.
Once you know the acoustic source positions, you know what you're dealing with. If the tweeter has to be moved forward a couple inches for alignment, then moving it back flush to the baffle would shift the vertical centerline upward. You might use asymmetrical slopes or staggered crossover points to shift it back. Higher orders cause more delay, and can be used provide apparent offset. Staggered crossover points change the phase between drivers too, and can be used to shift the null angles.
This is where measurement equipment really helps, because you can make adjustments and optimize the design much easier. The Smith and Larson
WTPro system is excellent for this, because it allows you to describe a crossover circuit using
Spice models, and measure the system with a digital representation of the filter functions described by the Spice model.
: ]
