To answer myself; I found the reason shortly after sending the mail. Two days 
of trial and failure :)

I set the spp value to the same as num bins, in this case 512 
(stream_args->args[«spp»] = 512). When changing this value to 508 the «ghost 
signal» disappeared, also with LO offset. I noticed earlier that the default 
max_num_samps() value reported was 508 when stream_arg was not set. So only 
question left is why spp should be 508, not 512?

If changing the bin size to 256, should spp be set to 252?

Fra: USRP-users <[email protected]> På vegne av Skorstad,Jørn 
via USRP-users
Sendt: tirsdag 22. oktober 2019 09:15
Til: [email protected]
Emne: [USRP-users] E310 LO offset problem

Hi list,

I am writing a C++ software which will scan through a given frequency range and 
perform an fft. The software seems to do what it should, however something 
weird happens when using LO offset tuning. With the following settings all 
looks correct:

Sample rate 6.4M - master clock 51.2M - 512 byte num_samples/binsize - offset 
0. However I can clearly see the DC peak for every «jump» in frequency.

With the same settings as above and a 3.2M offset there is a «ghost» signal 
3.2MHz below the real signal, which mirrors the real signal with slightly lower 
amplitude. This happens everywhere in the spectrum where solid signals (approx 
40-50 dBuV) are present. Using the command tune_request(freq, offset) and 
usrp->set_rx_freq(tune_request) to set frequencies. Have also tried with 
different gain settings from 0-40, no change. Have also tried with 10 dB 
attenuator between antenna and the radio, no change.

The same happens on other sample rates and offsets as well, the ghost signal is 
always the same as the offset (+ or -). I am using UHD_3.14.1. Have also tried 
changing rx_bandwidth without any change.

Is it possible to use LO offset and eliminate the «ghost» signals, or is this 
some radio/DSP limitation? I have pictures showing the problem clearly, I can 
send directly by email if necessary.

Thanks,
Jorn


_______________________________________________
USRP-users mailing list
[email protected]
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

Reply via email to