Still experimenting with multiple MaxBotix sonars trying to find a way to get rid of interference. Looking online it looks like this is a common problem, and even assembling all the sonars in daisy-chain mode, interference is still relevant: I tried both the commanded loop (you command the reading by strobeing a pin high, and then all sonars should read the distance in sequence, without overlapping) and the constant looping (you manually start the sequence and then the sonars read their distance in sequence, looping forever).
Finally, following the suggestions of the MaxBotix support, I tried commanding each sonar individually: and again, still the same result. The only difference is that this way I have more control over the intervals, so I can wait a bit more before sending another read sequence, and that with this individual triggering I can read the pulse pin instead of the analog value, and this is generally more precise: the only downside is that I use 4 pins instead of one.
But still readings are not stable enough: but at this point, since the duct tape test is ok, I think the problem lies in the double bouncing of sonic waves that goes around the room even after the measuring period has ended. But for that there is little to do electronically and from the programming standpoint: I’d have to try not to put mutually facing sonars, and trying to leave enough time for the waves to dissipate.