2122H vs 2202H for 200-800Hz

[4313B]

Recone your 2202H's to 2204H's. :dont-know

Way back in the old days when Project May first started up alot of people wanted to add a ten between the fifteen and the H9800. Greg was adamantly opposed to such a proposition. - "Won't work."

I think it is interesting that JBL uses dual eights with the Array horn in the SAM2LF though.

I know that the 1500AL and 1200FE both work really well. I could try my 2122H's or 2123H's at some point I suppose...

[ivica]

... beyond individual driver preference, it might be useful to also consider
in-system sensitivity matching, space requirements ("doghouse", baffle layout),
intended use (SPL goals/limitations), ... Although, starting with a nice
mid-bass driver and working out the rest to suit, strikes me as a reasonable
method as well.

Unless I already had other drivers in-hand, the 2122H is what I'd start looking to
get/assemble if finesse vs. slam were the intent.

It is expect-able that 2122 (as 'under-hung'), and 2123 and 2202 (as 'over-hung') would have some different character of reproduction, but here it is important that for each of them some "small" differences has to be made in network circuitry, in order to expect almost the same frequency transfer characteristics. In attached figures some simulations are presented under following :
-closed box of 15 Lit net-volume
-10W power

Presented graphs has been done under "in axes piston band response" presenting "piston beaming influence on higher frequency", and without that influence. On other figures network model is done with component values. JBL N3145 for mid-bass section was starting point for simulation.
Here in simulation it is expected to have one E120 rekoned with C8R2202 re-cone too, and marked as 2202H/E120. Stronger magnet on E120 basket produced some more efficiency in the mid-band region, and has to be attenuated more then 2202H. On the other hand some 'dumping' of "High-pass section" attenuated low-frequency response of 2202 driver with Rp1 resistor. This dumping, I expect that has to be done especially in the combination with 2245H driver as bass unit.
Rs1 is put nearer to input in order to rise the total box impedance in mid-bass region.

If no intervention done to N3145 (so Rs1=2.55R would be near Rp2=19.5R, and L-pad connected "in parallel with Rp2, and Rp1 removed, Cp2=14uF for all), but only L-pad attenuation to align sensitivity of the drivers third figure would be expected. As can be seen, 2202 i either version would have less bandwidth while 2123 would have larger than 2122.