Add a spectrum scanning function into the PlutoWeb UI

[unixpunk]

Should be pretty easy to take some values (freq range, gain, etc) from a new web page, stop services, run the scan, show the png file right there on the page.

[ImDroided]

This should be doable. As soon as I finish this basement in a few weeks I should have time to mess around

On Tue, Jan 8, 2019, 9:37 AM unixpunk <[email protected] wrote:

Should be pretty easy to take some values (freq range, gain, etc) from a new web page, stop services, run the scan, show the png file right there on the page.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/unixpunk/PlutoWeb/issues/40, or mute the thread https://github.com/notifications/unsubscribe-auth/ANVzaQnVDrxLT20sUgZl12jPVUOcRVQqks5vBLtJgaJpZM4Z1uBA .

[glangsto]

The wideband spectrum would be a great tool. How would you address the band pass gain issue, apparent in the ADI OSC program? The band pass gain sensitivity correction is needed to actually use the wide band spectrum.

Ascii output of intensity versus frequency would also be very valuable. Maybe just a link off of the web page.

Thanks

[unixpunk]

It wouldn't be a wideband view, it would be a sweep (like rtl_power) but LamaBleu has found some issue with rx_power which might delay this. I haven't investigated yet.

[LamaBleu]

Hi, discovering I'm not the only guy with this opinion about rtl_power. Quick wideband spectrum exists yet on Pluto, based on rtl_power. But only wideband : give usually very bad result on a "narrow" spectrum (let's say less than 500MHz wide).
Heatmap.py from keenerd will not work on the Pluto unless you have enough room on the flash to add needed modules.
But result can probably be improved using more than one pass. By using more than one pass rtl_power is able to create really huge files up to 1Gb, this can also create a new problem. However I will try to have a look on this next week. Sadly rtl_power_fftw (my preferred) can't be easily ported to Pluto. Perhaps we just have to rewrite a simple python or C tool to perform the same task.

[unixpunk]

You thinking a tool like hackrf_sweep except a Soapy_sweep or Pluto_sweep instead? Has anyone suggested any improvements to the rtl_power or rx_power to try to resolve it?

[LamaBleu]

Hi I took some time to have a look, so quick update about rx_tools on pluto!

• rx_sdr perfectly working (same as rtl_sdr) using CF32 format.
• rx_fm : working for WBFM mode. Still having issue on NFM, probably because the mimum samplerate is 200kS/s on pluto, perhaps we have to decimate. CF32 mode seems not available (TBC)
• rx_power : was able to make it run at 6MS/s instead of forced samplerate of 2MS/s. Hop-time is not a problem : 3 ms per 6MHz BW.. However I'm not satisfied at this moment by the result. Good but not accurate enough.

[unixpunk]

Thanks! For nbfm i think you still need to use rx_sdr and pipe to csdr because I don't think you can decimate via SoapySDR, it would have to be done in software, though not sure why having too high of a sample rate is an issue, I'll have to play with it and see if it has cmd line options for bandwidth since you can control the bandwidth in SoapySDR. (Maybe I'm off on all of this, just guessing)

for rx_power, I wonder if the accuracy is an issue because I'm compiling with -ffastmath and have speed over accuracy set for fftw in the fw itself. I might have to make another build with different options here if you don't find another avenue to improve accuracy. (I'm assuming accuracy isn't related to ppm drift but more signal gain-related)

[unixpunk]

Related to #50 ? Maybe a checkbox for the same screen that will enabled SNA mode (which enabled the TX portion) otherwise it just acts as a spectrum scanner?