Driver Selection on a high end 3 way

[harrisni]

Hi All,

I'm currently selling my B&W 803D floorstanding speakers to change to a DIY Speaker.

I've come up with 3 options.

Option 1
R2904-700009 SCAN-SPEAK 1" RING RADIATOR 2904 SERIES $395
12M4631G00 SCAN-SPEAK 2.5" CONE MIDRANGE $299
18W8531G00 SCAN-SPEAK 6.5" MID-WOOFER REVELATOR $259
18W8531G00 SCAN-SPEAK 6.5" MID-WOOFER REVELATOR $259

Option 2
R2904-700009 SCAN-SPEAK 1" RING RADIATOR 2904 SERIES $395
12M4631G00 SCAN-SPEAK 2.5" CONE MIDRANGE $299
21W8555-01 SCAN-SPEAK 8" WOOFER $239
21W8555-01 SCAN-SPEAK 8" WOOFER $239

Option 3
R2904-700009 SCAN-SPEAK 1" RING RADIATOR 2904 SERIES $395
12M4631G00 SCAN-SPEAK 2.5" CONE MIDRANGE $299
26W4867T00 SCAN-SPEAK 10" WOOFER - BLACK ALUMINIUM CONE $339
26W4867T00 SCAN-SPEAK 10" WOOFER - BLACK ALUMINIUM CONE $339


Option 1, I believe it will be alittle light on the bass side, but very very musical similar to the sonus faber designs.

Option 2, improved bass responce

Option 3, Again improved bass responce.

[harrisni]

I did find the B&W 803D alittle light on bass, even with 3x 8" bass drivers in each cabinet.

[augerpro]

Who's designing the XO?

[MadMutt]

http://www.troelsgravesen.dk/ektagrande.htm

Taken from here

http://www.troelsgravesen.dk/Diy_Lou...r_Projects.htm

[HiFiNutNut]

Looks like a very good plan.

Mine is Esotar T330D, SS18W8531G00 and SS26W8861T00. It is different from yours because mine is for OB, so the SS18W8531G00 would work much better than the 12M4631G00.

With your 3 options, In my humble opinions, I would prefer (2) over (1) over (3).

For your questions, I just spent some time looking at the datasheets. The 12M4631G00 appears to be a very nice midrange with a sensitivity of 90dB and a nice response. This means you can put the XO higher to avoid crossing at the most sensitive region of human hearing and with a shallower rolloff. Very nice. I would love it though I would be careful with its 4R impedance.

I would not cross the 4" 12M4631G00 lower than 500Hz 3rd or 4th order and if with a low order XO even higher XO point. I would do extensive modelling and simulations before deciding on what XO point at what order.

That means I would not choose option (3) with the 10" woofers because even with a 4th order XO I would still be worried about the imfluence of the cone break up of the woofers. Above 1k this woofer would not sound good at all, even at 500Hz it has some obvious response irregularity which would show up in a CSD as distortions. If I'd use this woofer I would probably cross it at or below 250Hz with a high order. Also, this is the most expensive option with 2 x 10" revelators.

With Option (2), the classic line 8" woofers look pretty fine up to above 2k. So if a 3rd order Butterworth is chosen at 500Hz, by 2kHz the response is 36dB down, which may be just sufficient. The woofer has a 87dB sensitivity. When using 2 woofers connected in parallel (4R load), 93dB sensitivity is perfect, if you consider the 90dB sensitivity of the midrange and some baffle step compensation. Of course, if you choose active XO, this may be irrelevant.

With Option (1), twin 6.5" 18W8531G00 in a bass reflex box should work very well too. You can even choose a 2nd order LR XO to cross to the midrange, which would be nice. The 8531 can be crossed higher than the 8" or 10" so if your XO is passive, you may be able use MKP caps and air core inductors only, which would sound more accurate.

However, note that in Option (2) you have at least doubled the cone area (SD) than option (1) which translates to much lower intermodulation distortions, deeper extension, and a far more relaxed bass and lower midrange. If you go with an active XO, you could choose a closed box for better transient response with a small amount of EQ. I would not do so with option (1) though.

In any case, since the midrange can't be crossed low, I would be careful with the driver placements. The distance between the midrange and the bottom woofer should not exceed the wave length of the XO point. That further explains why I would not choose option (3). I have not done the maths and it is only a rough estimte.

Given that the two woofers will be connected in parallel, the midrange is a 4R. You will have a very big challenge to design a fully passive XO. I would imagine the final XO has a very low impedance of less than 3R, or even 2R in a large region. You would need to build your own amplifier with a lower power rails but with a plenty of current reserves to do it justice, or probably a very powerful Krell may do it for you. If it were me, I would have the higher XO in passive, and the lower XO in active.

Regards,
Bill

[Brett]

Option 4:
JBL 2226 in about 100L port tuned to 40Hz
18Sound 10NDA610 in stuffed open back tube: JBL2123 is an option as would one or two PR170'd, but with a higher xover
Beyma CP380M on XT1086 waveguide.
All active: 300 (Audax 500), 1k5 xover points.

Will cost a little more than your options, but go lowish (still need sub for movie rumbles), loud and probably lower distortion. Bit bigger footprint though. Much more efficient.

[HiFiNutNut]

Brett,

I agree with many of what you said. Not sure if distortion can be reduced. From Zath's tests, the selected SS drivers have superb distortion figures that can hardly be matched by most pro drivers.

I guess your suggestion will be a whole lot easier to get it right and obtain good result. The risk of project failure would be a bit lower. But if for the SS, if very high power, high quality SS amps are used to conquer the low efficiency of the SS drivers, and if designed properly, the distortion figures will be very hard to beat.

I am open minded. Until I hear both I can't say which is a better option. I am confident that if both are designed well, both can sound top notch.

Regards,
Bill

[Brett]

Originally posted by HiFiNutNut
I agree with many of what you said. Not sure if distortion can be reduced. From Zath's tests, the selected SS drivers have superb distortion figures that can hardly be matched by most pro drivers.

Not so sure about that; where are the modern 10 and 15" pro drivers and compression drivers and flares tested on his site? At what level? JBL's (+Beyma and others) techdata and distortion show very good results for their drivers often at levels well above what the average person at home listens at.

Originally posted by HiFiNutNut I guess your suggestion will be a whole lot easier to get it right and obtain good result. The risk of project failure would be a bit lower. But if for the SS, if very high power, high quality SS amps are used to conquer the low efficiency of the SS drivers, and if designed properly, the distortion figures will be very hard to beat

Depends on your opinion of SS AB amps at low levels. Earl Geddes tests show something I've long suspected, ie that their distortion deteriorates at low levels.

Larger drivers need less movement to obtain a certain volume displacement (SPL) therefore, all else being equal, less distortion because of the motor. The usually very high power handling of pro drivers means power compression, usually fairly linear wrt Pmax will be much reduced and they usaully have a lot of attention paird to heat dissipation. I'm also of the feeling that heating of the coil is near instantaneous, therefore so is power compression, so throwing power at an inefficient driver brings other issues.

Because of weight issues, neo drivers are becoming nearly standard in new mobile speakers. Apart from the opportunity to redesign the motor assemblies to take advantage of this newish material, being smaller, it means a smaller magnet assembly with greatly reduced shadowing of the rear of the driver. Then you have a variety of other technologies like Differential drive (JBL), AIC (18Sound) and shorting rings that are at best rare in phile drivers

[ScottG]

Don't know what your level of experience is.. but you might want to take a look at these sites:

http://www.humblehomemadehifi.com/

http://www.troelsgravesen.dk/Diy_Lou...r_Projects.htm

http://www.linkwitzlab.com/

http://www.musicanddesign.com/

http://www.htguide.com/forum/forumdisplay.php4?f=39

[FullThrottleRic]

Originally posted by harrisni
Stuff

harrisni - I'm sure you have started a thread like this at least once before asking the same question, with the exact same options each time. Each previous thread (I'm sure there have been two) have given you the answers you're after