KerberosSDR sync issues

Discuss KerberosSDR - 4x Coherent RTL-SDR
Post Reply
mrice98
Posts: 5
Joined: Thu Feb 20, 2020 10:21 pm

KerberosSDR sync issues

Post by mrice98 » Thu Feb 20, 2020 10:27 pm

I have the FIR tap size set to 0, decimation set to 1, and antennas disconnected, but I am unable to get my KerberosSDR to sync correctly. In the image linked below, I have performed both sync steps (Calibrate IQ was pressed ~575 on the x axis in the bottom two graphs. In the ABS x Corr plot, I am not getting the three converging peaks. In the Phase Diff history, the red and green lines are converging but the yellow one is not. Does anyone know what could be causing this? I have tried restarting the software/my computer and switching cables and usb ports, but nothing has solved it. I am running the KerberosSDR directly from a laptop instead of using a Raspberry Pi or Tinkerboard

This issue did not occur until the first time I used the web interface for the first time, so I am wondering if it could be related to that.

Image: https://drive.google.com/file/d/17eH8SW ... sp=sharing

rtlsdrblog
Site Admin
Posts: 2677
Joined: Mon Nov 19, 2012 11:54 pm

Re: KerberosSDR sync issues

Post by rtlsdrblog » Fri Feb 21, 2020 1:07 pm

The web interface really shouldn't be related to this issue. But it looks like maybe some samples are being dropped.

Are there other programs running on your PC that could be hogging the CPU, and causing dropped USB samples?

And are you running the latest code? (Just updated a few days ago). The changes to the latest code ensure that the daq and sync C programs all run on the same core, which improves sample sync stability.

And I assume that this is not running in a VM? What are you laptop specs?

mrice98
Posts: 5
Joined: Thu Feb 20, 2020 10:21 pm

Re: KerberosSDR sync issues

Post by mrice98 » Sat Feb 22, 2020 1:28 am

I am running the latest code. I tried switching to higher quality USB cables, and it now is reaching zero phase difference with the peaks of the red and green lines in the ABS x Corr plot merging(however, the yellow line in the ABS x Corr is above the other two and does not have a peak). However, I am noticing that the plots go out of sync relatively quickly after attempting to use the DOA function for a bit. Is there something that would be causing it to go out of sync this quickly.

rtlsdrblog
Site Admin
Posts: 2677
Joined: Mon Nov 19, 2012 11:54 pm

Re: KerberosSDR sync issues

Post by rtlsdrblog » Mon Feb 24, 2020 1:58 am

Can you try running the ./kill.sh script in between attempts?

Can you check the spectrum plots in the first tab, and ensure that all are receiving the noise? (Compare the plots with the noise ticked and unticked).

And what are your PC specs?

mrice98
Posts: 5
Joined: Thu Feb 20, 2020 10:21 pm

Re: KerberosSDR sync issues

Post by mrice98 » Thu Feb 27, 2020 10:21 pm

Thank you, I will try that. The PC I am currently testing out the KerberosSDR on has an 8th gen intel i3 processor, 4GB DDR4 memory, and 128GB SSD storage. I am planning to later use it with a Jetson Nano (specs here: https://developer.nvidia.com/embedded/j ... eloper-kit)

mrice98
Posts: 5
Joined: Thu Feb 20, 2020 10:21 pm

Re: KerberosSDR sync issues

Post by mrice98 » Sun Mar 01, 2020 3:39 pm

I tried running the spectrum plots with and without noise, and it seems that the red and green plots are receiving noise, while the yellow is not. I am not sure what would be causing this.

Just curious- what do the different colors in the sync plots signify?

mrice98
Posts: 5
Joined: Thu Feb 20, 2020 10:21 pm

Re: KerberosSDR sync issues

Post by mrice98 » Sun Mar 01, 2020 4:08 pm

Update: I noticed that at 100 MHz, it will sync okay (yellow is slightly above red and green in ABS x Corr, all others are good). However, when I change the center frequency to 400 MHz and attempt to resync it, I experience the issues described above. Not sure why changing the center frequency would be causing the sync issue.

rtlsdrblog
Site Admin
Posts: 2677
Joined: Mon Nov 19, 2012 11:54 pm

Re: KerberosSDR sync issues

Post by rtlsdrblog » Mon Mar 02, 2020 1:39 am

It could be that one cable has become loose on the inside of the unit. Could you try opening it up, and checking if they are still connected? It could make sense, because at 100 MHz the noise could still couple through a broken connection.

Post Reply