Could someone send me an example screenshot of a QAM64 (or 128, 256) constellation diagram from a QAM analyzer connected directly to a QAM signal source? I need an close to ideal signal, but measured from real-life devices, not from a computer simulation.
There's plenty of ideal (i.e. theoretically simulated) constellation diagrams, where received symbols match perfectly a single point and there's a lot of real-life diagrams (from CATV meters, etc.) but there's no information if the constellation could look better.
They say, that this diagram has 0.9% EVM and "this value is representative of what might be seen in a headend or hub site, or the downstream output of a node."
My question arises from a QAM modulator and demodulator I designed - I'm not sure whether the BER and EVM I reached is the best I can get. I attached a QAM diagram of the signal after passing modulator (QAM128, 20 Msym/s), amplifier at 350 MHz and demodulator (including equalizer). So this signal has no impairments caused by the long transmission path. The constellation points are not too spread, but they aren't too compact either. Should I still try to improve this or is this best I can get?
Well,
your QAM looks ok. But to really evaluate, you need a full Channel model. i.e. the SNR level, BER performance,Symbol Synchronizer used...etc.There is also the important sampling frequency. It is difficult to say without the system spec whether it is the best or not.