I should have made some effort to format these numbers a little better, see if this is easier to read:
Behringer iNuke 3000 output power:
Load and freq Vrms P 8ohm P 4ohm
No load 1KHz 51,50 331,53 663,06
No load 20Hz 48,30 291,61 583,22
No load 14Hz 47,50 282,03 564,06
Max 8R 1KHz 51,00 325,13 650,25
Max 8R 20Hz 48,00 288,00 576,00
Max 8R 14Hz 47,60 283,22 566,44
Max 4R 1KHz 49,00 300,13 600,25
Max 4R 20Hz 47,40 280,85 561,69
Max 4R 14Hz 46,80 273,78 547,56
Measured by measuring output voltage, at the different frequencies, with different loads.
The P 8 and P 4 columns are calculated power output into respective loads if the amplifier outputs the measured voltage.
In very short intervals (i.e. peak transients) output power may be slightly higher than these static measurements indicate, but not by much, the output voltage is not likely to be higher than measured with no load.
Time interval of measurement is 2-4s burst.
Steady state output of very long intervals are not interesting for music/film use.
The amplifiers seems to have no problem running music or similar signals in to two 4 ohm subs, one on each output channel, running at max level.
I did not measure 2 channels simultaneously. If the psu is shared for both channels, the output power may be lower. To verify this, just check to see if there are two smps psu circuits inside.
Note that the iNuke has a built-in limiter that reduces the signal level in a way so that it will never clip the signal into square-waves like usual, so it is not immediately obvious by looking at the signal to determine that the maximum level is reached.
Lower output <<20Hz, as long as it does not freak out totally, is not a big problem as I see it, most drivers will be excursion limited at <<20Hz, so this may actually just help to make life a little easier for the driver.
This may be a limitation with inefficient, long-stroke drivers in small sealed boxes - the solution is of course to avoid such designs.
Also, keep in mind power is only 0.4dB down from 1KHz to 14Hz in 4ohm.