8945P + 27TBFC/G Floorstander

[sdclc126]

Hi all,

Love the new format - good job guys!

Here is my new "reference" speaker, which has been some time in the making. A few pics first, then some discussion (I hope). Usher 8945P + SEAS 27tBFC/G, with XO designed by Jay_WJ.

[sdclc126]

I didn't take build pictures as I felt there really wasn't anything particularly new to share, so I just took pics of the finished product.

The drivers were chosen based on Zaph's site for the tweeters and Jay's recommendation for the woofers. I started with the P.E. 0.38 cu. ft. cabinets which Jay used for a couple of his designs. I added a little bit of rib bracing to the interior side walls, then wall damping sheets, foam, and polyfill, all from P.E.

I prefer the authoritative look of floorstanders, but instead of just putting the cabs up on enclosed stands to "simulate" a floorstanding look or building floorstanding cabs from scratch, I had Home Depot cut some sheets of 3/4" mdf that I used to enclose the P.E. cabs on all sides except the front. This yielded walls of 1.5" thickness and take my word for it - the cabs are absolutely resonance free, at least audibly so, but the old knuckle test and the hand-on-the-cab-while-playing-music test yielded nothing perceptible at all.

The added width and height from the cabs prompted Jay to recommend slight XO mods from his original design for the P.E. cabs, and that is what is shown above. Dimensions are approximately 37.5 x 9.5 x 12 deep.

If there is anything unique about the cabs it is that they are essentially divided up into three sections - the top P.E. cab, an open middle section which as you can see houses the XOs, and another bottom enclosed section filled with sand. Everything below the P.E. cabs is double-walled to maintain the 1.5" thickness. These cabs are solid and heavy - about 80 lbs apiece.

Finish is very basic - textured flat black on the front and top to go with the baffles, which incidentally were wonderfully cut by Scott Stein at www.steinaudio.com - I have no router and no place to use one. The sides were first painted with Rustoleum hammered dark brown for a deliberate texture, then topped with a coat of dark brown satin enamel. No wood veneer for me - not right now anyway.

I put casters on the bottoms because we have carpet and I like to move the speakers farther apart from their normal positions next to the TV when I want to do serious listening. The casters are hardly visible because the cab floors are recessed. They are no spikes for sure, but it was a necessary compromise.

The speakers are supported at the bottom end by a powered Peerless 8" XLS (no longer made) and they blend together very well; I actually run them fullrange, but the woofers are attenuated at the bottom by a single 250 uF cap - this leaves me some flexibility with subwoofer XO and level adjustments.

As for the sound, well it is simply the best and cleanest I've ever heard, and I can't recommend this very high value driver combination enough. I fully expect that these will be my primary speakers for years to come, however when I have a larger, dedicated listening space I will probably look into OBs, line arrays, and other designs, but perhaps use the same drivers.

Thanks again to Jay_WJ for his input and guidance on this build.

Well that's about it - any questions or comments?

[MisterTwister]

Very nice. I think it was right choise to use Jay's crossover. Usher woofer is properly filtered. Unlike some other designs where they push woofers up to 2-3k

[pkitt]

I, too, built some speakers using the Usher 8945P. They are MTMs with a BG Neo3PDR planar tweeter with the woofers working in mass-loaded transmission lines. Dan Neubecker designed their elegantly simple crossovers having just 8 components; 3 inductors, 3 capacitors and 2 resistors. I modeled the ML-TL lines using Martin King's Mathcad-based TL worksheets and named them Marcato. Their modeled F3 is 32 Hz. The crossover is installed upside down in the base.

[sdclc126]

PKITT - what is the crossover point in that design? I believe the BG needs a relatively high XO point.

P.S. - very nice build.

[pkitt]

The crossover point is ~2300 Hz and worked out quite nicely sound-wise. Around that frequency, both woofers and the tweeter are rolling off at nominal 2nd-order rates, but at about 2 kHz the tweeter's rolloff becomes close to 8th-order while the woofers continue at 2nd-order. Thanks for compliment on my build.

Originally Posted by sdclc126

PKITT - what is the crossover point in that design? I believe the BG needs a relatively high XO point.

P.S. - very nice build.

[sdclc126]

Originally Posted by MisterTwister

Very nice. I think it was right choise to use Jay's crossover. Usher woofer is properly filtered. Unlike some other designs where they push woofers up to 2-3k

Thanks. Regarding the 2k + XO point, see PKITT's comments in Post #6; I'm not sure if the Usher's response bump is addressed there - PKITT - can you comment?

[pkitt]

There is about a 3-dB dip in the Usher's response at about 1.3-1.4 kHz which was just partially mitigated by the woofer's response already starting to rolloff a bit below 2 kHz, but otherwise we just lived with it. I can't hear that dip's presence and I'd be surprised if anyone else could, either.
Paul

Originally Posted by sdclc126

Thanks. Regarding the 2k + XO point, see PKITT's comments in Post #6; I'm not sure if the Usher's response bump is addressed there - PKITT - can you comment?

[Jay_WJ]

Originally Posted by pkitt

The crossover point is ~2300 Hz and worked out quite nicely sound-wise. Around that frequency, both woofers and the tweeter are rolling off at nominal 2nd-order rates, but at about 2 kHz the tweeter's rolloff becomes close to 8th-order while the woofers continue at 2nd-order. Thanks for compliment on my build.

Paul, you post here, too!

I liked Dan's approach to designing this crossover for the driver combination. With these extremely asymmetric slopes, It will mimic the smooth power response of LR2 above the xover point.

MisterTwister,

After many tweaks and experiments with the Usher 8945P, I concluded that crossing it up to 2.5 kHz is okay if the breakup node is notched out properly, which is what Dan did for Paul's MTM. But there's one problem if one does not have a measurement setup. By measuring several units, I found that the 8945P's breakup node changes up to +/- 300 Hz due to its unique cone---QC must not be easy with this type of cone. I wish Usher made a poly cone version of the 8945P Poly or mineral-filled poly cones are not so popular these days. These are excellent cone materials IMO except that they can look cheap.

Softdome, excellent build!

-jAy