[USRP-users] Questions on USRP N200 data streaming
Ralph A. Schmid, dk5ras
ralph at schmid.xxx
Thu Sep 18 04:30:34 EDT 2014
> But in the case of SDR, all the bits and pieces (including the software bits in
> your signal flow-graph) are uncalibrated until you actually calibrate
Anyway there is one thing, that usually needs not very much calibration, if any - that is the frequency. Not only by means of the carrier frequency, but also the bandwidth of the emission. Even a lousy crystal is enough, that for example the deviation of a SDR generated FM signal is on the point at the predefined value, symmetrical and precise.
Voltages and signal levels can fool you quite a lot, like always with analog stuff, but the precision of SDR generated waveforms is impressing and hardly can be reached with conventional technology, at least not with this versatility, conventional solutions usually can only handle a very limited range of waveforms. This is because at SDRs the mathematical theory directly calculates and produces the signal in 100% predictable quality, instead of simulating the maths with some analog stuff (mixers, amps, modulators, filters) that is tweaked and twisted until it fits somehow.
For example, when I need to realign a conventional FM receiver, I take one of my $$$ SDRs, let gnuradio spit out a 145/433/whatever MHz signal, modulated with for example 1000 Hz at 2.4 KHz deviation, and the IF filters and the demodulator can be adjusted with an accuracy that in former times only could be achieved with bulky radio testers that were paid $$$$$. Not to talk about digital stuff, where I have every single bit under my control with gnuradio.
> Marcus Leech
More information about the USRP-users