AnsweredAssumed Answered

Expected outputs to go with example flow graphs

Question asked by Exray on Jul 1, 2016
Latest reply on Jul 30, 2016 by Exray

I am trying to learn about GNURadio and provide support to colleagues with less time and computer experience for an upcoming project using two AD9625's with the goal of streaming the output to a linux PC for graphing/processing. I am not an RF expert and need to come up to speed quite a bit.  I'd like to confirm the test setups have working drivers/programs and better understand making successful connections so that problems doesn't confuse setup issues with otherwise working hardware.


What I have available:

ZC702+FMCOMMS5 with latest SD image from ADI wiki instructions

Linux PC running GNU Radio Live SDR with persistence, with compiled in support for libiio, libad9361-iio and gr-iio. I can see/add FMComms stuff.

(In a week or more, the goal is to replace the FMCOMMS5 with our board with the 2xAD9625, which sounds like we'll need to make a custom block for in place of the FMCOMMS5 sink/source blocks).


So other than reading the tutorials (and more), running the examples seem like good way to get a better handle on how things work and tweak from a working setup to see how changes affect the output. But I can't seem to find the expected output other than what is used in the tutorials I've come across. I can see from reading this forum that some of the examples are out of date and would/may not work anymore as-is.  Knowing what the output should be would confirm whether things are working as expected, if there is something impairing the setup, or just using the wrong values.


I tried to copy and modify a flow graph I saw on this forum from a person having driver issues that took in a sine, triangle, square and saw tooth waveform and am able to build it, but the outputs displayed do not match my expectations, which is clean output of the sine, triangle, square and saw tooth waveforms. 


So then I went to the simpler cyclic-sine example, expecting a nice clean sine wave being output.  However, by default it looks like noise until I change from 100ns/div to 25us/div, and then it somewhat resembles a sine wave, but still looks bad.  It does look like it might be 32.768k based on the 25us/div and eyeballing it.


Screenshot 2016-06-30 20.14.29.png


Here is what the flow graph looks like for cyclic-sine.grc:


Screenshot 2016-06-30 20.16.25.png



On the linux PC, where I was trying to do 4 waveforms:

Screenshot 2016-06-30 20.40.53.png



1. Does the output from the cyclic-sine.grc look correct? I recall seeing another thread on the forum that the output is expected not to look like source unless modulation was used, or something to that effect.

2. Is there anything super obvious about what is wrong with the 4 waveforms flow graph that would explain why its not working and just looks like noise?  (I need to get a bigger monitor so the graphs can be displayed on screen for screenshot and also get speakers to do the audio examples).

3. Do I need to have loopback cables installed on the FMCOMMS5 board? I don't see it mentioned anywhere but I was thinking I do and that the Sink block takes in I/Q data over the FMC and converts it to analog going out the Tx SMA connections and then the FMComms5 source takes it from the Rx SMA connectors and converts it back to I/Q data for processing on the linux PC.  So far, I've only seen loopback cables mentioned for the production diagnostic testing, so not sure if the sink/source happens internally to the AD9361's and never involves the SMA connections.  Since I think I see the 32.768k sine above without loopback cables, I've been operating without them until I find out otherwise. Unless its bad because that is an open air connection being made instead of cabled in with lots of signal...


I appreciate any response for this newbie. I learn a little more each day, and its not lost on me that I have a long way ahead of me to get a better grasp of this system.


Thanks in advance!