New SB Gema DIY 3-way speaker design

[Rick Sykora]

Was perusing the SB Acoustics website for drivers and came across a new 3-way speaker design:

Gema - Sbacoustics

Open Source Kit: Not available as full kit, read below for details. Gema is a large 3-way speaker paying homage to the classic speaker of the 1970's. Generous enclosure proportions fitted with large bass drivers were the bee's knees at the time. And for good reason! The Gema is our take on a...

sbacoustics.com

Appears aimed at the market for those pining for the days of big woofers with wide baffle cabinets....

The drivers are all pretty reasonable however it uses a high part-count crossover that will drive the price a bit higher. Gema is marketed as part of SB's open-source effort so is more of a bare-bones approach vs their other more complete DIY kits. Nice to see support of DIY from a reputable supplier of contemporary speaker drivers.

 

[ryanosaur]

Interesting.
Surprised they chose a 4" mid instead of something a little larger just to compete with that giant 12" Woof in displacement. Would be nice if they provided some sample measurements to show efficacy of design as a whole. The Drivers all look solid and seem a good match with each other.
17 element XO would hopefully yield a very flat response, and if voiced well be neutral and accurate.

 

[thewas]

Surprised they chose a 4" mid instead of something a little larger just to compete with that giant 12" Woof in displacement.

Crossed at 680 Hz the woofer will reach its limits earlier with most music, also since the 1" tweeter has no waveguide 4" is better than something smaller from directivity point of view.
Good design which I would like to listen and built myself.

 

[ryanosaur]

Crossed at 680 Hz the woofer will reach its limits earlier with most music, also since the 1" tweeter has no waveguide 4" is better than something smaller from directivity point of view.
Good design which I would like to listen and built myself.

Confused about the Mid... I was thinking that a 5 or 6" would be a better mid considering the Woofer is 12". Obviously still making certain it matches well with the Tweeter, too.
Thoughts/clarification, please?

Definitely agree with the closing statement!

 

[thewas]

From SPL point of view excursion increases 4-fold(!) with each octave lower if you don't change driver diameter or to keep the excursion similar you have to double(!) the driver diameter for each lower octave. From 640 Hz to 40 Hz there are 4 octaves so you can calculate how much bigger the woofer would have to be compared to the mid driver. In reality mid drivers usually don't do as much displacement as our hearing is more sensitive to distortion in the mids so the difference is smaller but nevertheless crossed at that frequency even a good 2" or 3" mid wouldn't be problem from SPL point of view.

From directivity point of view you don't want big discontinuities at the crossover points and a driver larger than 4" beams already quite at 3 kHz so a 2"-4" mid is a better partner to a 1" non-waveguided tweeter crossed at 3 kHz. A simple approximation of such is comparing the diameters of the driver with the corresponding wavelengths, for example the directivity of an ideal pistonic 4" driver at 1 kHz would similar to the one of an ideal 1" at 4 kHz as the corresponding wavelength is 4 times smaller.

 

[ryanosaur]

From SPL point of view excursion increases 4-fold(!) with each octave lower if you don't change driver diameter or to keep the excursion similar you have to double(!) the driver diameter for each lower octave. From 640 Hz to 40 Hz there are 4 octaves so you can calculate how much bigger the woofer would have to be compared to the mid driver. In reality mid drivers usually don't do as much displacement as our hearing is more sensitive to distortion in the mids so the difference is smaller but nevertheless crossed at that frequency even a good 2" or 3" mid wouldn't be problem from SPL point of view.

From directivity point of view you don't want big discontinuities at the crossover points and a driver larger than 4" beams already quite at 3 kHz so a 2"-4" mid is a better partner to a 1" non-waveguided tweeter crossed at 3 kHz. A simple approximation of such is comparing the diameters of the driver with the corresponding wavelengths, for example the directivity of an ideal pistonic 4" driver at 1 kHz would similar to the one of an ideal 1" at 4 kHz as the corresponding wavelength is 4 times smaller.

Cool. Appreciate your explanation!

 

[Rick Sykora]

Not really fond of a larger speaker like this needing a 6 degree tilt when up on a stand.

Suggest would be better to have a cabinet design with a slanted front baffle to address.

 

[somebodyelse]

I suppose the cabinet is complex enough that you could assume the builder is up to the task of a slanted baffle, but making a stand with tilt and support at the back is a lot easier.

 

[ryanosaur]

Not really fond of a larger speaker like this needing a 6 degree tilt when up on a stand.

Suggest would be better to have a cabinet design with a slanted front baffle to address.

Is the cant recommendation due to a tilt in the crossover dispersion? Or is it just to mimic the other Speakers of said "vintage-esque" design where they are lower to the ground and tipped up and back to play toward ears of the listener?

 

[Rick Sykora]

Is the cant recommendation due to a tilt in the crossover dispersion? Or is it just to mimic the other Speakers of said "vintage-esque" design where they are lower to the ground and tipped up and back to play toward ears of the listener?

Potentially, 3-way makes it more challenging to alter lobing and may cause other response issues.

Despite the vintage aspect, I would want the best contemporary design from a fancy crossover.

 

[alex-z]

They have a 20mH inductor as part of a series notch filter on the woofer. I think it would be smarter to pick a different driver that doesn't need that treatment.

 

[Rick Sykora]

They have a 20mH inductor as part of a series notch filter on the woofer. I think it would be smarter to pick a different driver that doesn't need that treatment.

Along with the really big cap, this filter is more a contour than trying to fix major woofer misbehavior.

Perhaps taking a bit too much design freedom when not worried about selling many, I guess. 

 

[Schollaudio]

Along with the really big cap, this filter is more a contour than trying to fix major woofer misbehavior.

Perhaps taking a bit too much design freedom when not worried about selling many, I guess. 

I don't think it needs that kind of compensation especially for BR. Some guys might want that in a sealed box.

 

[Rick Sykora]

I don't think it needs that kind of compensation especially for BR. Some guys might want that in a sealed box.

Agree, that filter does seem out of place. However, without asking SB or simulating the entire design, hard to know exactly what the designer intended.

 

[Wolf]

Typically, it compensates the self resonance of the woofer when values are that large. In the bass response range when using low DCR coils in series, a high magnitude woofer resonance can react with the lowpass coil and produce a considerable peak in response around resonance. This is one of 4 ways to deal with those kind of results. I'm certain I've listed them here at ASR prior in another thread about kit designs and effective measures.

In this case, the designer has listed Icore coils with 1.4mm diameter wire. Those are not going to have much resistance to damp the bump inherent. Also of note is the 1.2 ohm damping resistor in the shunt of this 3rd order electrical circuit. This will pull the knee down in the lowpass rolloff more than without, and likely exacerbate the look and magnitude of the inherent bump. Without that resistor, the hill might not be as severe.

I would have likely tried a 50 ohm 25W resistor across the woofer to see if that did the trick more inexpensively and not cut much sensitivity. Increasing the lowpass coil DCR likely would kill too much sensitivity of the woofer. A parallel L||C(series R) may do the trick also, but then result and cost is about the same.

 

[Kwesi]

Wolf already explained the use of the series shunt parallel to the woofer, without there will probably be a hump at 80~100Hz in the FR. It is pretty common and also often needed in 3-way designs also for the midrange driver e.g., look at Troels Gravesen or German Hobby Hifi Magazine. The remaining parts are pretty straight forward electrical 3rd order design, probably to achieve acoustical LR4 slopes.

17 element XO would hopefully yield a very flat response, and if voiced well be neutral and accurate.

I would not bet on this. E.g., this is FR of SB Acoustics Ara kit as offered and sold:

there might be a reason why they do not publish measurements on their website....

Bernd Timmermanns had to fix that:

Improved Kit:

SB Acoustics ARA-Kit

Kompaktlautsprecher,SB Acoustics ARA-Kit,Bassreflexbox,Lautsprecher,Bausatz,SB Acoustics Satori TW29R-B(sb-satori-tw29r-b), SB Acoustics Satori MW16P-4(sb-SATORI-MW16p-4)

www-lautsprechershop-de.translate.goog

SB Acousics builds nice drivers, but I would wait until indepentend measurements of the Gema are available...