Some files missing

[Box88]

Thank you for the quick reply. I tried to build gr-doppler again but some files are missing in the lib directory: CoordTopocentric.h, CoordGeodetic.h, Observer.h and SGP4.h.

[darkstar007]

OK. I'd forgotten about the fact I link with a external library that provides me with the satellite prediction functions. Unfortunately the way their Makefile was written means that you can't just simply install it, but have to tweak the Makefile a little. I'm a bit pressed for time this evening, but will add some instructions tomorrow about which is the library to use and what tweaks you need to do to the Makefile (from what I remember its just adding a '-fPIC' to the CFLAGS).

[darkstar007]

Sorry for the delay - I've added some simple notes in the file 'INSTALL' at the top level. If they work for you I'll close this issue.

[Box88]

Thank you again! I installed everything as described in the install.txt file. Now my problem is that when I drag the doppler_c block from the blocks list to the flowgraph sheet, GNU Radio Companion crashes. I've already installed gr-ec blocks and they work but I cannot figure out what's wrong with doppler_c block.

[darkstar007]

Hmmm. I've never tried using the Doppler block in GRC - I've only ever used it in hand coded python scripts. I'll have a look into this. This may take slightly longer....

[darkstar007]

I've not had time to test it, but at least now we stand half a chance of it working! The XML file which GRC read was just the default one for this project - I've now modified it so it may now work..... let me know how you get on. There's a bunch more things I need to fix - like changing the altitude into metres from its current kilometres.

[Box88]

OK now I can run the Doppler block in GRC (I added the frequency and sampling_freq inputs which are missing in the block). But I get wrong results, in terms of Doppler shift frequency. So I read the c++ file doppler_c_impl.cc and I don't understand how the shift frequency is computed in the 'work' member function (especially in the for loop). Could you explain it, please?
BTW I started looking at the python files in the src folder and it looks a nice job!

[darkstar007]

Hi. Thanks for pointing out I was using the wrong prototype in the GRC xml file.

The thing to know about this block is that it provides a phase correction not a frequency correction. This means that in order to use it you take your (complex) signal of interest & connect it to a 'multiply' block, then connect this block to the other input of the 'multiply' block - and the output signal of the 'multiply' block is now Doppler corrected.

The work function computes the doppler shift at the start and the end of the block of work its been given to do (seems to be around 4096 samples typically). I then linear interpolate between these two values for each of the samples and compute the phase correction necessary - this approximation will be fine if the sampling rate is high.