The Napoleons: a Compact 2 Way that punches above its weight


2004-09-22 3:19 am
The Napoleons are a small DIY loudspeaker designed to deliver a natural tonal balance and clean loud playback from of a small box suited to both small and large rooms, at a reasonable price. This write up was in response to requests from the 2013 Ottawa DIY fest, where the speaker was first shown

Driver Selection
Off axis response irregularities will detract from speaker tonal balance more in a small room than in a large room due to the greater relative strength of indirect sound, compared to direct sound. To help combat this, a tweeter was chosen with some mild horn loading, and a woofer with a well-controlled extended higher frequency response that could take advantage of it.

The Tang band W4-1720 (Tang Band W4-1720 4" Underhung Midbass Driver 264-872) provides powerful and clean mid to upper bass from a very small sealed box. A vented design was avoided to keep the box small and to avoid box/vent unloading which would compromise the high sound pressure levels desired for larger rooms. The mighty little W4-1720 plays ridiculously loud and clean in this application, and mid to upper bass is rich enough to give a pleasant tonal balance. It’s smooth mid treble allows a higher crossover point so the tweeter can play loud without strain. This is a good sounding driver, but it can still be made to bottom when over driven by drum test discs. Still, it plays clean and loud enough to pass the “I didn’t realize it was playing that loud” test.

The tweeter is the Vifa D26NC55-06. It’s well-controlled off axis response near crossover and its drooping high frequencies off axis reduce “splatter” in a small reflective room. Its distortion is also very low. While this driver is officially NLA, examples can often still be purchased from on-line suppliers.

Box Dimensions
The sealed box is simple and was built using hand tools and a flush trim bit. Half inch mdf kept the box small which in turn keeps panel vibration under control. The box was made shallow to allow true shelf mounting. Woofer center is 3.5” from bottom, tweeter 7.5”. Drivers are centered horizontally.

A different baffle shape (for example due to thicker wood) would change diffraction and affect the response. This could require a crossover change and this design is expressed for 0.5” wood only.

Crossover and Frequency Responses
The crossover is a 6 element second order electrical, in phase driver connection. Special attention was placed on smooth on and off axis response, and a smooth response when standing while listening.
Low pass
L1: 1.5 mH inductor, 0.7 ohms dc resistance (18 gauge air core).
C1: 15 uF cap (not electrolytic)
R1: 1 ohm power resistor (10W or greater)

High Pass
C2: 2 uF cap (not electrolytic)
L2: 1.1 mH inductor, 0.5 ohms dc resistance (18 gauge air core)
R2: 15 ohm resistor, high power (25W or greater)

The full project write up is attached and includes system box tuning, material plan, numerous measurements from 20 Hz to 20kHz, along with design rationale and more detailed build information.

Dave Dal Farra


  • 1.jpg
    49.7 KB · Views: 2,868
  • 2.jpg
    32.9 KB · Views: 2,774
  • 3.JPG
    33.3 KB · Views: 2,722
  • Napoleons.pdf
    664.8 KB · Views: 422
  • 4.JPG
    63 KB · Views: 2,716
Last edited:


2010-05-04 5:02 pm
Nice. I built a quite similar speaker not too long ago using the old Genesis tweeter with that woofer. 2kHz acoustic crossover point and almost the same driver slopes too, but a bit more recessed upper midrange and a little downward tilt (I was voicing it for a room that's basically a glass box, so it needed to be extra mellow). I don't know if I'd call the woofer ridiculously loud running without a high pass, but it is surprising how good it sounds at ridiculously high excursion. Oops, just realized you're running it sealed, so that's probably +6dB or so over what I had with a lot of music.
Last edited:
Re: Dave's comment #6; "surprisingly more perceived bass than indicated by the curve".

Could we look at this in a bit more detail pls?
This perception turns up occasionally in listening tests and it is a very interesting reaction. I recall in Martin Collom's book "High Performance Loudspeakers" (originally about the late 70's and several editions since) that he reported an 8 litre (internal volume) two-way (probably the little BBC monitor) came out best amongst a large group of speakers, most of which would have had greater internal volume with bigger bass units.
He writes of it as follows: "...(it) was found capable of sounding natural in a "full" and "spacious" manner---more so in fact than many of the larger systems to which this quality is more likely to be attributed."

Immediately following this observation he summarizes the test of the 30 sets of speakers in this way.

"Subsequent analysis revealed that the favoured speakers where those which possessed very even axial response over 100Hz to 10kHz when measured by third octave and octave averaging. Mid-band uniformity and overall spectral balance would thus appear to be a crucial area of speaker performance."

Now Dave would seem to have achieved those latter goals in his design and is reaping the benefits. My edition of Collom's book is 35 years old now. Phase in the mid-band is just one of the items receiving more attention since then. Are there other facts that people are aware of...stumbled across etc. It would be helpful if we could identify those so as to be able to intentionally reproduce them in subsequent designs.
In another thread recently someone observed good bass could come from small enclosures and he thought it depended on the "attack". Which may be an accurate subjectively observation but I'd like to know more about what we can do to generate that.....presumably it was system "Q"......but I am not certain.

I would be interested in any contribution that might expand/clarify/explain this phenomenon. (i.e. of smallish speakers subjectively "boxing above their weight") Think of the benefits. Smaller speakers means less cost, less work, less hassles domestically......all good!
Congrats Dave.

Cheers, Jonathan
Last edited:


2004-09-22 3:19 am
Hello again

The frequency responses shown in my earlier post were crossover simulations using real quasi-anechoic driver in-box measurements.

I measured the system response tonight just to verify the accuracy of the simulations. As seen attached, they're dead on. I've also updated the pdf with this information.



  • Napoleon Sim Verification.JPG
    Napoleon Sim Verification.JPG
    45.3 KB · Views: 928
  • Napoleons.pdf
    683.9 KB · Views: 83


2004-09-22 3:19 am
Thanks Jonathan.

I believe the additional perceived bass is also an artifact of the dip in response between 250 and 500 Hz as well. Apparently this was a little trick used by the BBC on the LS35A, see Figure 4:
BBC LS3/5a loudspeaker 1989 Measurements |

Large speakers and a lot of room treatment don't fit with my biases on what a pleasant domestic environment should look and feel like. Driver selection, box and crossover were purposely designed to be small but try and sound bigger and naturally balanced in a reflective environment.

I think a second order high pass at 80Hz and a 500W small sealed sub would make a dandy and unimposing system.

These also have the small box imaging thing happening as well. I really do like this W4-1720 driver.




2004-09-22 3:19 am
There are several ported designs available but what I tried to distinguish with this design is the use of a sealed box, for smaller volume and the ability to play louder than ported, given the high tuning. I think this presents a unique set of trade offs.

I can't really comment on use of any other replacement tweeter unfortunately. This tweeter was chosen for its somewhat unique (narrower) dispersion characteristics, for a dome. I think a tweeter that has a more typical wider dispersion will sound bright with this tuning, and would need the top end tilted down a bit.


PS Dave, good to meet you too, hope your project works out