The cross-post makes it hard to know where to respond but, since my comments are in a similar vein to the previous poster's, I'll add to this thread.
While any spectrum analyser is better than none in many circumstances, the poor resolution bandwidth is also a pretty severe limitation of the TinySA in many applications. Two-tone audio-spaced tests and diagnostics of digital transmissions are two examples.
I agree that a tracking generator is useful but I've found myself doing more with the nanoVNA recently, partly because the resolution bandwidth of my Rigol RSA3015E-TG go-to unit is quite poor (1 kHz) in TG applications such as crystal filter characterization. I don't recall that limitation in my earlier DSA-series unit but it seems to be a characteristic of the 3000N series, too. Using the nanoVNA-F with NanoVNASaver worked better for my phase-noise measuring notch filters, although I did find a bug in the plotting routine when used on narrow sweeps. Anyway, the developers fixed that, and I'm grateful.
You get what you pay for, although the bangs per buck are undoubtedly highest at the entry end of the field. On the other hand, having too many limitations gets pretty wearing during any serious work.
73, Peter.