On Fri, Sep 27, 2019 at 03:44 AM, hwalker wrote:
I don't understand all the theoretical limitations associated with different
FFT sizes; but from a practical use point of view if my $50 device, as a
additional throw-in, can measure a 65 meter length cable accurate to 0.1 meter
then I'm impressed.
more FFT size leads to better time resolution. I don't need to measure cable length with TDR, this is not reliable way, because it depends on cable velocity factor which may vary even through cable length and is not stable, it very depends on temperature, frequency and other factors. More easy and reliable way to measure cable length is to use meter or with measurement tape.
TDR allows to analyze cable and connector issues, such as bad contact, impedance change, delay to incident plane, etc. But it needs at least picosecond resolution to be usable. This is why firmware TDR implementation is useless. With nanosecond resolution it's just a toy for those who want to see what TDR is. NanoVNA memory is too small for better resolution, because it needs at least 16384 points FFT. But you can do it on the PC.
I think that waste valuable controller memory for useless feature such as TDR just for one-time use DEMO purposes is a bad way. It is better to use this memory for more useful features, such as better calibration, more measurement points, more graphs etc. If you're needs TDR, you can do it on PC with much better resolution
