My questions are:
Which drivers need the BSC?
How to measure the “z-offset” and which ones need it?
I am working on a set of three way tower speakers using the ScanSpeak Discovery Drivers. I am using the following drivers:
ScanSpeak Discovery D2608/9130 1" Textile Dome HDS Tweeter.
Scanspeak 18W/8434G-00 Discovery, 7" Midwoofer (18 cm).
Scanspeak 22W/8534G00 Discovery 8" Woofer (22 cm).
I am taking inspiration (and some liberty) from Troels Gravesen’s “ScanSpeak Discovery 861.” That is the reason for the drivers I am using and the basis for the cabinets I intend to make. My twist is that I am going to put an “Epique E180HE-44 7" DVC MMAG Extended Range Subwoofer” in each tower in their own enclosures at the base. I have already built a subwoofer using one of these drivers and was impressed with the output. The subwoofers will be ported and run together powered by an external amplifier so the crossover will not need to provide signal for the Epiques.
I have started building a test box as a full scale mockup of the speakers to test the configuration. The woofer will have its own ported chamber (above the sub), and the mid woofer and tweeter will be in an enclosed chamber at the top of the cabinet. I am going to experiment with having the subwoofer forward facing and downward facing (the subwoofer I made using the Epique sub produced the best sound in my room when it was downward firing). The baffle is in two pieces for the test box, that way I can swap out baffles to test subwoofer configuration.
The cabinets will be ¾ inch plywood and the baffle will be one inch thick plywood. The subs get two inch round ports, and the woofer gets a three inch round port.
I am asking for crossover design help. I realize that to really tune the speakers properly, I need to get frd and zma files from the actual drivers and I plan on doing that as soon as I get the baffles made and the holes for the drivers cut out. I will run frequency sweeps in the actual test box. The woofer, mid and tweeter will all be on the same common baffle and I will flush mount them. And counter to what people recommend, I will line them all up on the same axis for symmetry (aesthetics). For now I’m using the frequency responses and impedance from the manufacturer’s datasheet as my test data.
I am looking for advice on the three way crossover. I have done one two-way and now I need help with the three way crossover. Specifically, with “baffle step correction (BSC)” and “z-offset.”
Any input would be greatly appreciated.
Cheers,
George, Tigard Oregon, UDA
Which drivers need the BSC?
How to measure the “z-offset” and which ones need it?
I am working on a set of three way tower speakers using the ScanSpeak Discovery Drivers. I am using the following drivers:
ScanSpeak Discovery D2608/9130 1" Textile Dome HDS Tweeter.
Scanspeak 18W/8434G-00 Discovery, 7" Midwoofer (18 cm).
Scanspeak 22W/8534G00 Discovery 8" Woofer (22 cm).
I am taking inspiration (and some liberty) from Troels Gravesen’s “ScanSpeak Discovery 861.” That is the reason for the drivers I am using and the basis for the cabinets I intend to make. My twist is that I am going to put an “Epique E180HE-44 7" DVC MMAG Extended Range Subwoofer” in each tower in their own enclosures at the base. I have already built a subwoofer using one of these drivers and was impressed with the output. The subwoofers will be ported and run together powered by an external amplifier so the crossover will not need to provide signal for the Epiques.
I have started building a test box as a full scale mockup of the speakers to test the configuration. The woofer will have its own ported chamber (above the sub), and the mid woofer and tweeter will be in an enclosed chamber at the top of the cabinet. I am going to experiment with having the subwoofer forward facing and downward facing (the subwoofer I made using the Epique sub produced the best sound in my room when it was downward firing). The baffle is in two pieces for the test box, that way I can swap out baffles to test subwoofer configuration.
The cabinets will be ¾ inch plywood and the baffle will be one inch thick plywood. The subs get two inch round ports, and the woofer gets a three inch round port.
I am asking for crossover design help. I realize that to really tune the speakers properly, I need to get frd and zma files from the actual drivers and I plan on doing that as soon as I get the baffles made and the holes for the drivers cut out. I will run frequency sweeps in the actual test box. The woofer, mid and tweeter will all be on the same common baffle and I will flush mount them. And counter to what people recommend, I will line them all up on the same axis for symmetry (aesthetics). For now I’m using the frequency responses and impedance from the manufacturer’s datasheet as my test data.
I am looking for advice on the three way crossover. I have done one two-way and now I need help with the three way crossover. Specifically, with “baffle step correction (BSC)” and “z-offset.”
Any input would be greatly appreciated.
Cheers,
George, Tigard Oregon, UDA
All of them, technically. However you shouldn’t need to worry about it if you have measurements.
All of them need it but you don’t have to worry about it if you use the right measurement technique.
AllnB raises another question for me. I understand that you can get the TS parameters by using something like the Dayton DATS (which I have). I am now wondering if putting the drivers in the cabinet is a "good" way to get the impedance sweep data for the zma file and if that also is a "good" way to generate the frequency response (frd) for each driver to then use in Xsim or one of the other software packages? (I have Xsim and Vituix, though I prefer Xsim).
To get true high fidelity, you should find a way to make all the magnets in the 3-way cabinet align vertically. The subwoofer won't matter so much. One easy method is to design the cabinet so the baffle is tilted back a little at the top, instead of having separate baffles for each driver, so the magnets are aligned vertically that way (it won't matter if the drivers are angled upwards, since the angle won't be much, and a listener will still be in a sweet spot out front). But if you would rather use a crossover to compensate for all drivers on the same baffle, as you said, you must provide a delayed signal for the mid and tweeter. However, including passive time-delay circuitry in a crossover is problematic, at best. You will be better off tilting the baffle. Even some high-end speaker manufacturers use that technique. Otherwise, don't worry about that and go with your gut; performing as many trial-and-error tests with crossover component-values as you can to get the kind of response you're looking for.
For a great reference, I recommend the book "Loudspeaker Design Cookbook" by Vance Dickason (KCK Media Corp., first published in 2006). In it you will find all the information you need to build a high-quality speaker cabinet, along with extensive instruction on crossover design.
For a great reference, I recommend the book "Loudspeaker Design Cookbook" by Vance Dickason (KCK Media Corp., first published in 2006). In it you will find all the information you need to build a high-quality speaker cabinet, along with extensive instruction on crossover design.
Last edited:
Thank you Kurtus. The tilting method is new to me. I do have an idea for pyramid shaped speakers that this tilting of the baffle is part of the design. As for the current project, I'm going for an easier cabinet to assemble as my carpentry skills are a bit lacking.
I do have a version of Dickason's book. I have found it is hard to understand and I seem to get more confused each time I read through it. But it did teach me a lot of the terminology used.
I do have a version of Dickason's book. I have found it is hard to understand and I seem to get more confused each time I read through it. But it did teach me a lot of the terminology used.
Compensating delay in a passive crossover is not difficult. While a delay network could be used, most opt to adjust the filters themselves for a little more rolloff to create the wanted effect.
Agreed. As long as the highest possible fidelity is not a requirement, such as with studio monitors. But for ordinary listening, that tactic will work fine. However, since GeoWorx DIY has a copy of Dickason's book, I still recommend that he takes the time to thoroughly study the section on crossovers.