Failing to receive IQ samples in gnu radio

I have done a fresh install on my linux machine of the heimdall daq, doa repository, gnu radio and gr-krakensdr. Running the doa code works as expected, the kraken connects and I can browse waterfalls in the browser. However, when I attempt to use gnu radio (after changing daq_chain_config.ini output from shmem to eth, and starting heimdall with ./heimdall_only_start.sh) I repeatedly receive the error “Exception: Failed to get IQ samples”, even with extremely simple flowgraphs and those from the examples directory. It does show some reconfiguration success messages, and when I put the krakensdr block into debug mode, the sync_state changes around, so it looks like some connection is made, but no IQ data.

I will collect the gr-krakensdr debug output later this evening to share - are there any other debugging steps I can take, or obvious causes? Thank you kindly for any assistance.

The main thing to check is the log files in the heimdall log directory. Can you check and see if there is anything interesting in those?

The gr source block will only capture “DATA” type frames and ignore “DUMMY” or “CALIBRATION” type frames from the DAQ IQ server. My guess (based on a similar experience) is that the DAQ is still in calibration mode and therefore the calibration data that the DAQ sends across the ethernet interface is probably being dropped by the gr source block (as expected). The gr source block will report “Failed to get IQ Samples” when the read packet queue is empty which can be the case if the DAQ is taking a good amount of time to get into calibrate at startup. I’ve found it takes about 30s to several minutes some time for calibration but its dependent on the DAQ configuration settings and your particular situation.

Unfortunately, the gr source block doesn’t give one real time feedback as to the calibration state. I’d recommend starting the DAQ Firmware, let it run for several minutes, and then try to run the gr-source block to see if you start seeing IQ data. It might take several attempts at restarting the gr-source block.

I’m curious if anyone else has experienced something similar.