Hi Lee,
IF the satellites are reasonably flat to a frequency that is "significantly" below the chosen crossover point, and IF the bass panels are reasonably flat to a frequency that is "significantly" above the chosen crossover point, then you want to have the SAME 3db bandwidth point for both the high pass and low pass filters. What is referred to as the crossover frequency would correspond to that 3db bandwidth point, and at that point each filter will have rolled off by 3db relative to the response further into its passband.
So if you choose both "corner frequencies," as they are also called, to be exactly the same, and if the slopes are the same, then the overall response will sum to flat.
When I used the word "significantly" in the first paragraph, how significant that is will depend on the sharpness of the filter slope.
That is all a different situation than in the design of INTERNAL crossovers within speakers, where the choices of corner frequencies and slopes may differ between the high pass and low pass parts of the crossover, in order to take into account differing driver characteristics near the crossover frequency.
Assuming the criteria of the first paragraph are applicable, if you use a passive high pass filter, it seems to me that you also need to use a 6db/octave (first order) slope on the low pass active filter, or a mismatch will occur.
If you implement both the high-pass and low-pass filters actively, then you would have the flexibility to choose both slopes to be anything you want, within the range of selections that are offered by the electronic crossover.
My understanding is that while very sharp slopes may be beneficial in keeping the input frequencies to each speaker within the speaker's comfort area, the downside may be increased artifacts in the crossover, such as passband ripple in the frequency response, and phase anomalies. Those are complex areas that I'm not particularly familiar with; perhaps someone else will provide additional insight.
Best regards,
-- Al