Recommendation for 2.5way SB Acoustics

Hi guys!

A friend of mine has two beautiful cabinets, pretty much the same size as the sb acoustics Arya kit.
SB Acoustics :: Arya Kit

As we both are completely noobs with crossover design, we decided that the best we could do is to buy the kit without the cabinets, since the cabinets are identical (almost same baffle size, almost same liters) and we like the sound of the sb acoustics.
The bad thing is that the frequency response at the arya kit is very lame (see the attached picture)

I know that the cabinet, the position of the drivers and other stuff play important role, but my friend is not a perfectionist so, we are just looking for a better crossover design than the sb acoustics kit.
The drivers that we intent to use are sb acoustics sb29rdc and sb17nrx or Sb17mfc


  • Screenshot_20200108_233924.jpg
    400.7 KB · Views: 709
Last edited:

Hi vassilis1984,

I've highlighted which part of the frequency response is affected by the room which part does the crossover effect. The crossover cannot effect the response that the room did.
The crossover is very good by interpreting how close the 3 measurements are to each other from midrange and up (~500hz - 20khz) and when they start to part from one another, how smooth they deviate.

This chart is perhaps too honest and have indeed scared less experienced people shopping for a kit.

I also own the Ara and Rinjani and both are fantastic.
BTW, the freq resp looks OK, maybe I would tune the BR lower - longer port. Because the ring dome has off axis response falling down off axis they try to compensate by raising top octave. You can try very small inductor 0.05mH in series with the tweeters to flatten it. Can you post FR with drivers out off phase - let see what the crossover do in time alignment.
The only criticism about the crossover I can make is that the response rises to +5dB at ~15kHz. Looks like that can be partially compensated for by not 'toeing in' the speakers, such that they aren't directly aimed at the listener.

All the raggedness below 1kHz is a room/measurement artifact. The woofers are operating as pistons in that frequency range so the anechoic response will be dead flat down to 100Hz, below which the response will roll off.
I have not listened to this 3-driver TMM Arya design / Kit.

From my experience with the 6" paper and aluminum cone SB17xxx35-8 midranges, the Arya's 3,000Hz cross-over frequency is too high for the best sound since the 6" driver is already beaming and at the beginning frequency range of cone break-up. Several popular SB17xxx35-8 designs use a 1,650-1,850Hz crossover to avoid midrange cone breakup, and some favor a BW3 tweeter circuit to avoid generating sibilance from tweeter resonance effects from a shallow-slope crossover. To avoid tweeter siblance, designers using a stepped baffle for tweeter-midrange time alignment select a 1,850 - 2,000Hz crossover with the required +LR2 and -LR2 crossover circuits.
The Arya cabinet is 195mm = 7.7" wide.
If you had an 11" wide cabinet you could construct a 3-way T-M-W design using an 8" woofer.

SB17 midrange frame diameter = 171mm 6.7"
SB23 woofer frame diameter = 234mm 9.2"
8" woofer has 63mm 2.5" wider frame

The Smart Kids use this extra 2.5" -3" baffle width around just the tweeter and midrange to reduce baffle edge distortion by cutting Avalon style truncated pyramid bevels, or 1" - 1.5" radius quarter-rounds. If your amplifier can easily drive a 3-ohm load, using a 4-ohm woofer and an 8-ohm midrange produces enough SPL/watt difference for a modest baffle step compensation. Using higher resistance wire inductors on the midrange and very low resistance iron core inductors on the woofer can also modestly help SPL balance with 8-ohm midrange and woofer.

SB17NACS34-8 has Fs=31.5Hz and 86.5db @ 2.83V @ 1m

SB23NACS45-4 has Fs=23Hz and 90db @ 2.83V @ 1m
SB23NACS45-8 has Fs=25Hz and 87.5db @ 2.83V @ 1m

.....keep searching and comparing alternative proven designs before you purchase
The SB17NRXs are problematic above 2.5k, that's for sure, I'm kind of surprised that people at SB used 3k as x-o point because it turns into Himalayas past 3.5k ! �� and breaks up around 6k - something fishy.
With 5-8deg sloped baffke it's also easier to time align the drivers and the tweeters would be a smoother. I struggled with x-o to keep it around 2k LR4 (keep tweeters distortions low) and attenuate the Himalayas so they are low in level and don't interact too much in low-mid highs with tweeters. The fronts are 1.5" with large round over to keep difraction low with tweeters 1" of the center line and lower Q somewhere between 24dB/oct LR and Bessel (29RDNC).
Line Source how do you like Alu cone SB17 comparing with the paper and same for tweeters ALu or CDC vs fabrics. Personally I don't think the neo ringdome should be crossed lower than 3k as they looses transparency and starts to sound harsh. Also my pair fs is in 800-900Hz - way above the listed in datasheet. Or it could be the cone break up from the 17NRXC, but it's down 20dB at 4k and 30 dB at 6.5k.
Last edited: