I’m doing a subject on wireless comms as part of a uni course. Not much of it makes sense to me so far, though I’ve managed to muddle through as best I can.
But Quadrature Amplitude Modulation. Urgh. Just the name makes me want to brain myself with a corkscrew. Can anyone explain this to me, like I’m an idiot? I “get” the bit about changing phase, but that’s where it ends.
These constellation diagrams don’t make any sense to me at all. What are they supposed to represent?
I have an assignment question that tells me to convert the first few letters of my name to ASCII (easy), which I have then converted to binary (also easy). Then it says to encode it using QAM. I have no idea where to start on this one. Can someone give me a clue, please?
Think of those ones and zeroes actually being sent down a phone line. How, exactly? All you can do at the other end is pick up some voltage level. So, you could just say any voltage below a certain level = 0, above it = 1. “1010” would then look like a square waveform, with the voltage jumping between levels.
However, this isn’t very efficient. It would be more efficient (=higher bandwidth) to encode more than one 1 or 0 for each change in the voltage waveform, say, four at once. So, if we had two voltage signals, then one signal having low amplitude and zero phase compared to the other could be, say, 0000 and a signal having high amplitude and 90[sup]o[/sup] phase could be, say, 1111, with all other permutations (1010 etc.) being different phases and/or amplitudes. (They are usually plotted with their “in phase” component versus their out-of-phase ‘quadrature’ component, not just phase versus amplitude.) The constellation shows the particular strings of 1’s and 0’s you choose for certain phase and amplitude combinations - there’s plenty to choose from, but the higher the bandwidth you seek, the more likely the receiver will make a mistake in identifying a string and thus more errors will creep in.
Regarding explaining the constellation diagram - imagine thinking of the amplitude and phase of a signal at some instant in time plotted on a graph in polar coordinates. It would be a dot, right? Now imagine watching that dot move around as you progress through time in slow-motion. In QAM, that dot would go near one corner point, hang out there, then drift over towards another corner point, hang out near there, then go towards another corner. If you drew the motion of this dot, it would be kind of a mess, but you could see that it would converge mostly around the corners and you’d see traces of how it gets from one to another.
This is basically a constellation diagram, but usually they’re drawn not with all of the traces, but they show just where the dot is at the signal’s sample points, all overlaid on each other. There will usually be a smear of samples near each corner, with good signal quality indicated by small smears near the ideal corner dots, and poor signal quality by larger blobs instead. If the blobs get too big, any dot could be interpreted as being near the wrong corner, which would be an error. The distance from a sample dot in that smear, to the ideal placement, is called the error vector (if it’s a vector you’re talking about) or just its distance is the error vector magnitude, or EVM, which is a very important parameter in complex modulation.
Ok, I think I’m understanding the idea of QAM now, but the constellation diagram still has me confused. What makes a particular pattern “rectangular” vs “circular”? I mean, I can see that in the picture, but is that the only reason it’s referred to in that manner?
And still, how do I apply the constellation diagram meaningfully? I’m trying to imagine a sine wave that stops and starts all over the place and it seems a bit odd. Or is that the whole point?
The sine wave doesn’t stop and start all over the place - it’s always there, but its amplitude and phase are always varying. A good-quality signal will sit close to one of those ideal dots for most of the sample period, then quickly transition to a different amplitude and phase for the next sample. The sine wave is always going, though, and typically on a constellation diagram you just see one point for each sample period, taken in the middle where the signal is presumably closest (and where the receiver will be sampling). These are just snapshots, all overlaid so you can get a statistical view of whether the real signal was close to the ideal point on the amplitude-phase diagram, at the sample time.
Most QAM is a 4x4 rectangle, but there are some unusual cases of the ideal points laid out on a circle. If you are designing your own modulation scheme, you could place the ideal points anywhere you want - circle, rectangle, pentagram, you could have it outline a Mickey Mouse face for example - but in the real world that would be hard to achieve, and having equally-spaced dots is best.
