5 Attachment(s)
3-way with 2235H, 2123H and SEOS-12
I've built a pair of cabinets with curved sides and the aforementioned components. All panels except the sides are made from 3/4" MDF; the baffle is two 3/4" MDF panels laminated together. Each side was made from 6 laminated layers of 1/8" hardboard, and required no kerf cuts. I CNC-machined the baffles on both sides: rebates, dadoes and flush-mounts for all drivers. I will be rounding over the cabinet corners with a 3/4" bit when they are next back in the shop.
Cabinets are 44" tall, 18" wide (front), 12" (back) and 15" deep. 44" places the waveguide at, or just slightly above, ear-level when seated in any number of desk chairs, which is what I usually have handy, and also allowed the 2235H to be raised a bit higher off of the floor. Distance between each driver is 1.25". Sealed cabinet of approximately 4.8 cu ft for the 2235H after subtracting bracing and separate boxes for 2123H and SEOS-12. I chose to isolate the waveguide from the back-pressure waves of the 2235H in order to minimize the chance that they could stress the waveguide walls. The compression driver is a Denovo Audio DNA-360.
Cabinet is fully stuffed with fiberglass and braced with 2X4's.
I have very little experience in the acoustical/electrical parts of building speakers. I would like these to be a fully active setup, preferably with as little complexity as possible. I find the capability of software crossovers very intriguing, in particular the combination of Equalizer-APO and rePhase. When combined with an HDMI connection that can carry 6 separate channels of audio, I would only need one A/V 7.1 Receiver to do all of the amp work. I have listened to this combination of drivers using an Ashly XR-4001 and separate Yamaha amps, and liked it very much; however, I would like to learn how to do a proper crossover design rather than depend on my ears.
I would very much value input into several things:
- Measurements: impedance and gated impulse response, particularly in as much detail as possible. (as in, I need an English-language description of how to set up amps, mic preamps, software, cabling etc.) I do not own a measurement system (LEAP, Omnimic, CLIO or similar), and so planned to use either HOLMImpulse or Room Equalization Wizard along with a mic preamp.
- Physical placement of microphone relative to the front baffle when executing impulse response measurements - distance from baffle, and x/y location relative to baffle, especially the y-axis for the waveguide and 2123H.
- Measuring individual drivers versus measuring combinations of drivers with the same signal:
- Why would I choose to measure drivers individually, versus together? Is there a good book or series of articles that anyone can refer me to?
- If I measure all drivers with the same signal, how do I prevent damage to the compression driver from receiving low-frequency signals?
- If I measure each driver independently, how do I account for the vertical offset of drivers (I believe VituixCAD has the ability to incorporate these offsets; I'm learning how to use that software).
- Level-matching between all three drivers:
- I assume that it is better to pad down to the 2235H, rather than up to the compression driver.
- Matching levels may be more functionality in VituixCAD than I am currently aware of. If anyone on the forum has experience with this software, any pointers would be greatly appreciated.
- I will likely be forced to measure these in a small room, which will probably limit my usable far-field impulse response to about 400hz. I will probably need to merge both near- and far-field measurements for the 2235H in order to get valid low-frequency data.
- Anyone here have experience performing these two types of measurements?
- I plan to crossover the 2235H at about 250 hz, where both it and the 2123H should still be radiating omni-directionally. I assume that I need not be concerned about a directivity match between these two drivers, but must absolutely be concerned about it between the 2123H and SEOS. From the polar plots I have seen of the 2123H and SEOS, it appears that approximately 1700Hz would be appropriate.
- The 250Hz frequency is basically arbitrary. The 2235H seems to have some "fatness" at frequencies above 250Hz, although I am aware that the 4430/4435 used it up to 1000Hz. My subjective observations here likely are biased due to room effects, poor level-match between the 2123H and 2235H, and who-knows-what-else. Is there a solid reason why I would raise this crossover frequency?
- Crossover slopes
- With linear-phase filters, where (to my knowledge) I need not be concerned about the phase-shifts inherent in minimum-phase crossover typologies, I have never found a convincing argument to be made favoring a steeper crossover slope over a shallower one (or vice versa), other than to reduce the amount of sound being broadcast by more than one driver simultaneously, usually for power-handling reasons. Any insights, and/or reference articles?
I know this is a lot, so If you've read this far, I both commend you and appreciate it! I am aware that the SEOS guys over on diyaudio and avsforum have done quite a bit of crossover work with this waveguide, and I'll likely be asking over there for help also.
A few photos of the build:
Attachment 79500Attachment 79501Attachment 79502Attachment 79503Attachment 79504
