Avoid is kinda strong. There are good reasons for using narrower filters in a multicoupler, particularly as new cell phone usage starts up in the 600 to 800 Mhz bands. Your post pretty well gives the reasons for narrower filters. The 411A does a better job than most of front end filtering but that advantage is somewhat over ridden by any multicoupler, more so by wide ones with weak amplifiers. In a well designed multicoupler not only will the filters be only as wide as necessary but the internal amp will be low distortion at high input levels in order to not produce RF intermod products. This spec is commonly left out by some manufactures, as it is hard (expensive) to accomplish. Instead they will quote amazing noise figures which are easy (read cheap). Ideally the amp in a multicoupler will have low gain, low noise, and low distortion (a high third order input intermod number). As usual with things RF, these desirable traits are not easy to attain simultaneously. A really strong, low intermod RF amp can make up for wide band input filters, but the ideal is narrow filters and strong amps.
Watch out for quotes of output intermod number. These values are always higher than the input intermod value and make for better numbers. What really measures the performance in a multicoupler is the input intermod value. A high gain, low power amplifier can have good output numbers but weak input numbers since the input number is the output value MINUS the amplifier gain, i.e., high gain leads to poor input intermod values but usually excellent noise figures.
One way to improve the performance of a wideband multicopler, is to use antennas with built in filtering or inherent narrow band response. For instance, an SNA600 dipole has about a 30 MHz bandwidth. That is equivalent to having a 30 MHz filter at the input of the multicoupler. A Yagi antenna would be an even narrower bandwidth. Sharkfins (log periodics) have wide response so are not good "filters". Powered sharkfins with built in filters can help.
Another way to protect a wideband unit is to put a low loss inline filter in front of the wideband multicoupler input and then swap out the inline filters depending on what bands you are operating in. As an easy example, the Lectro PF25 is a one block wide passive filter and the PF50 is two blocks wide.
What the user would like to have is a wideband antenna system and a wideband multicoupler that does not introduce spurious signals (low intermod) and is usable for all possible wireless frequencies. As in most RF compromise, as the airwaves become more congested this dream is going to become a little bit of a nightmare or at least a nightpony.
Best Regards,
Larry Fisher
All very true, though I would say, if it has an amp at all, then input intermod values need to be considered. I would like to see PSC measure and publish third order input intermod numbers rather than just an excellent noise figure. See discussion above.
Best Regards,
Larry Fisher