[matevana]

Just completed the first prototype of a design I have been working on for a few months now. It's a true dipole 3-way, measuring a mere 15" x 23". The highs/upper mids are handled by a ScanSpeak 10F/8424G (which is surprisingly pleasant sounding way up top) crossed passively to an Usher 8137A 8" driver, at about 1800Hz, 12dB/oct. Bass duty is handled by a Dayton RS 12" driver, using a seperate power amp with build in xo at 120Hz.

Initial impression is quite promising. A slight peak centered at 630Hz, perhaps some floor bounce, is currently being tamed actively until I can implement a proper parallel notch into the design.

The Usher, with its fixed offset, and Scanspeak, with its adjustable swing offset (pivots left to right) really do have an affect on diffraction issued measured off axis. Initial testing shows the best compromise is achieved with the ScanSpeak locked in it's left most position, as shown in the picture below. This combination measured better than the test baffle, with both the Usher and ScanSpeak in the 12 o'clock position. The second speaker will be a mirror image of the first. More to come...       

[oskar]

Very nice build there! Will they go on stands? How far down will they reach? Is the xover separate or mounted onboard. What is the sensitivity? Is there a brace around back?
Great stuff, love to hear them.

[matevana]

Very nice build there! Will they go on stands? How far down will they reach? Is the xover separate or mounted onboard. What is the sensitivity? Is there a brace around back?
Great stuff, love to hear them.

The 'stands' are built in, and suspend the panel on 4 (squishy) sorbothane pads. They do a great job with dampening any panel/floor interaction. The design incorporates some floor bounce and the panels sit at a slight upward angle, about 18 degrees. They perform strong down to low 40's and generally taper off rapidly by the low 30's. Panel resonance is surprisingly low, due in part to the small baffle/large driver combination. The drivers are clamped through with 8-32 x 1 1/4" cap screws which adds to the overall rigidity of the system. So far I see no need to add additional bracing, but did build a brace if need be. The onboard crossover is board mounted and sits next to each mid-woofer, suspended by 4 rubber faucet washers. The x/o only addresses the mid/high interaction and also provides 3.5dB attenuation for the scan speak driver. The overall system sensitivity seems to be around 86db. The presentation is very authentic. The ScanSpeak driver is truly amazing and has lots of detail throughout its operating range. The Usher 8137 is no slouch either and this combination seems to work well together. If anything, the Dayton RS seems to be the weakest link in the chain; I previously used a 10" Peerless XXLS and think I preferred the low detail, but the Daytons still need to break in. The surrounds are very stiff at this time.   

[Bob in St. Louis]

Nice, well done! 
Those would be great flanking a large PC monitor. Granted, a bit overkill.... but ..... 

Bob

[matevana]

Nice, well done! 
Those would be great flanking a large PC monitor. Granted, a bit overkill.... but ..... 

Bob

That's how I'll be using them Bob! The picture makes them look bigger then they really are, but they are actually smaller than many floor-standers, and only a few inches deep. I cut a smaller matching baffle for an open back center speaker, using only the ScanSpeak and Usher drivers, via the "small" speaker setting for HT duty. The center will measure a mere 12" x 12" and will have the ScanSpeak pivot from the top as well.  It will sit on a console table just under my plasma TV. Not overkill at all. The cost for adding the center channel was $138 for the drivers and a few bucks in parts. I have yet to hear better SQ from any center costing < $200.

[Bob in St. Louis]

Well that sounds pretty cool.
If possible, please post pics of the final setup. Sounds neat. 

Bob

[nickd]

Could the peak at 630hz be caused by a baffle reflection? you might try some careful placement of a foam or felt waveguide (front and or back) before adding a notch filter. Notch filters always seem to drag a little life out of those fast, transparent drivers IMOP.

VERY nice work by the way. I wish I was that creative.

[matevana]

Could the peak at 630hz be caused by a baffle reflection? you might try some careful placement of a foam or felt waveguide (front and or back) before adding a notch filter. Notch filters always seem to drag a little life out of those fast, transparent drivers IMOP.

VERY nice work by the way. I wish I was that creative.

Thanks Nick. In looking at the Usher's response curve, there is a 3-4dB hump at those frequencies. Panel (or even floor) reflections may be compounding the situation, as you suggest. I agree with you about the affect on SQ while using a notch. They often require large coils with significant DCR. Right now I'm using a DBX 31 band EQ and it takes adjustments to only 3 bands (5-6dB) @ 500Hz, 630Hz and 800Hz to get the system response basically flat. This unit adds no noise to the system so I may choose to keep it instead. It also addresses the slight raggedness of the ScanSpeak's response at the highest octave.   

[Bill Baker]

Interesting design and looks very nice by the way. What kind of low end are you getting from the Dayton driver in this arrangement?

[matevana]

Interesting design and looks very nice by the way. What kind of low end are you getting from the Dayton driver in this arrangement?

  Thanks for your comment. It seems quite promising.  I say this because I have built one panel so far, and am working on the mate. The Dayton HF (as opposed to HO) uses a thinner aluminum cone which seems to preserve more detail in the lower octaves. For now however, the surround is very stiff and needs some break-in. The motor seems to be properly vented, and I hear no mechanical noises. Also, xmax is fairly high at almost 15mm, one way. I really do prefer flat panels to other alignments such as H and W, and am encouraged by the Daytons ability to keep up without much help. 

[poseidonsvoice]

Wow, once configured properly, that would be a state of the art PC speaker no doubt!

Beautiful! Care to share some of your measurements?

Anand.

[matevana]

Wow, once configured properly, that would be a state of the art PC speaker no doubt!

Beautiful! Care to share some of your measurements?

Anand.

Anand, I dont have any measurement software, just my trusty calibrated mic and a frequency analyzer, so let me describe what I am seeing. It's an all active system, so I EQ to get the in-room response basically flat, measured somewhere between on-axis and 15 off center. 30 degrees off is pretty respectable but I tend not to give it much consideration as my listening position is pretty much fixed.  The woofers do a good job keeping up mid to low 40's. By low 30's, they're pretty much done. Without any EQ, there is a 6dB peak centered at 630Hz, which is down by 3dB on either side at 500 and 800. The ScanSpeak mid/high driver is basically flat over 5-6 octaves, which is pretty amazing. I add +3dB overall starting at 15kHz up, and then make small adjustments to smooth out the SS slightly ragged response, from 15kHz up. Once flat, I like to dial in a slight "U" shaped house curve, which to me sounds best with 2-3dB boost centered at 80Hz, and another 1-2dB centered at 4k. It uses a wide Q, so the taper is very gradual and basically flat in the more critical midrange.