From The RadioReference Wiki
Current generation digital scanners, even upgraded to the most recent firmware (where applicable), do a poor job of handling simulcast transmissions on any system using P25 CAI. This is most often noticed by a breaking up of a fairly strong transmission that one would normally expect to be crystal clear. This problem is most evident when monitoring true Project 25 simulcast trunking systems (9600 baud control channel) using the LSM CQPSK modulation scheme, however it can also affect monitoring of digital voice talkgroups on Motorola SmartZone system (3600 baud control channel) sites operating in simulcast configuration. In this article, we'll describe and explain the most common reason for this sort of problem, and suggest some ways you might alleviate it if it troubles you.
This problem is generally due to the scanner receiving multipath signals and not being able to properly correct the digital signal so that the audio can be recovered.
Multipath reception describes the situation where your receiver receives the signal from one or more transmit sites over paths of different lengths. Generally, single-site multipath isn't a problem; the FM capture effect will pick the strongest signal, as multipath reflections will be of lower power, or the wrong polarization. But in multi-site simulcast trunking, there are multiple transmitters on separate towers, and the power from more than one of them can be close enough for your receiver to pick them both up.
When receiving multipath signals from an analog system you may hear just a bit of wavering in the signal and your ear/brain has no trouble detecting the correct sounds coming from the speaker. (Shortwave signals that suffer this cause more trouble in single sideband (SSB) due to the nature of the way the signal is handled and those of you who have experienced listening to multipath SSB are well aware of how tiresome it is to listen to.) In a digital signal, the vocoder in the scanner must detect every bit (0 or 1) that is coming in, assemble these bits into a packet then translate this stream into an audio signal that is understandable by us. The problem we're discussing occurs because the signals which come in from separate towers are not in time synchronization when they get to your scanner, because the paths are different lengths. As long as there is sufficient overlap in the bit positions, a good receiver can extract the audio in the face of this multipath interference, but when the slip is more than one bit time, it results in a broken transmission, usually resulting in what a lot of folks call a "pixelated" signal. It usually manifests itself as a readable signal followed by a broken signal then back to a readable one, etc.
If the above situation was present in real radios (the ones actually used by the subscribers), then P25 CAI (indeed any digital signaling) would not be acceptable for a public safety use. It should be pointed out that this multipath distortion is much less prevalent in real radios but in fact this does happen a little to real radios. You may hear some user actually comment about it when asking someone to repeat something that "broke up" when you heard the requested repeated transmission just fine. In addition, the system engineers have the advantage that they can "tweak" the actual system's timing values until the problem is all but eliminated.
Since we, the scanner hobbyist, are not paying really big bucks for a system that must work properly, the error correction algorithms used in the scanner to deal with this problem are not nearly as strong or as good as those used in real radios. This is not particularly the manufacturer's fault but rather a situational problem. To properly come up with a better algorithm for error correction, the software engineer must be able to "tweak" the code by repeated trial and error. And once it works for system 'A', will it work as well for system 'B'? Also, if the specific system that you are listening to has not been "tweaked" to the exact right specifications, this will just exacerbate the problem for the scanner user. Remember, actual systems can be adjusted properly or nearly properly and that one variable can cause us problems.
There is no "one" solution to this problem; however, there are things that you the scanner user can do to mitigate the problem depending on your situation.
- See this post by UPMan: Loose/Open Squelch Improves LSM Simulcast Reception
- If you are closer to one site than another, you may not experience this problem at all. This explains why in discussions about the problem on specific systems, some people find it intolerable while others hardly notice it at all. This leads to one solution for base station monitors; a directional antenna pointed at the site you want to monitor. If you are receiving all of the signal from only one site, there is no multipath distortion to deal with. This of course does no good for people who are mobile.
- Attenuation of all the signals sometimes helps. This is of course due to the fact that if you attenuate the signals you possibly lose the ability to hear the interfering signal from the multipath source.
- In lieu of attentuation, remember that the *less* gain your antenna system has, the better off you'll be; reducing the overall signal at the receiver input will (generally) increase the signal-to-signal ratio, so capture effect will ameliorate the problem for you.
- Keep your firmware updated in your scanner. The scanner manufacturers are indeed addressing this problem as more and more of these simulcast digital systems come on line. Of course sometimes it may seem that the most current version of firmware takes a step backwards, but due to the fact that each system is unique, it may well be that worked better on system 'A' actually causes system 'B' users to notice a backward step.
An important thing to remember is that this problem is much more likely to happen to scanner listeners than actual users of the system, because -- at least in the best case -- all the components of the actual trunking system have been engineered and set up so that the levels at the edges of each site's RF pattern are either timed properly to avoid the problem, or the power levels are chosen to reduce it as much as possible, or... and there are probably 4 or 5 different things an engineer can do within a system to reduce simulcast multipath issues... but they're not doing them for *scanner listeners*.
This article about HDTV has some additional information about multipath and other distortion / interference problems. The discussion "maps" well to public safety multipath issues, as well.
One individual had several issues with this wiki article as it existed on Jan. 2013. Rather than attempting to rewrite the article, the reader is referred to 
Return to the Uniden DMA FAQ