[USRP-users] RFNoC Block with 2 inputs and 1 output

Dario Pennisi dario at iptronix.com
Sat Oct 14 04:50:31 EDT 2017


You may want to check this:https://www.mail-archive.com/usrp-users@lists.ettus.com/msg03271.html

Dario Pennisi








On Sat, Oct 14, 2017 at 10:31 AM +0200, "Andrew Thommesen" <andrewjohn83 at outlook.com<mailto:andrewjohn83 at outlook.com>> wrote:


Hi Dario,


Thanks. I did try that, but I am using a TwinRx daughtercard that runs at a slower 100MSPS so I don't think its an issue. It does work when I connect the inputs to radio and outputs to host blocks. My problems start when I try to connect one of the outputs to a radio for Tx from a WBX daughtercard......


Andy
________________________________
From: Dario Pennisi <dario at iptronix.com>
Sent: 13 October 2017 19:32:33
To: usrp-users at lists.ettus.com; Andrew Thommesen
Subject: Re: [USRP-users] RFNoC Block with 2 inputs and 1 output

Hi Andy,
As I wrote you have to check available bandwidth. For x310 input data is 32 bits at 160 MHz whereas radios run by default at 200 MHz so you won't be able to fit 800 mbyte/sec in a port able to receive just 640 mbyte/sec.
Note that setting sampling frequency in rfnoc radio is useless. To change sampling clock you have to set system clock attribute in device3. Of course same thing goes for outputs....

Doing this worked for me... I'm connecting inputs to radio and outputs to host blocks...

Let me know if this helps...

Dario Pennisi





On Fri, Oct 13, 2017 at 9:15 PM +0200, "Andrew Thommesen" <andrewjohn83 at outlook.com<mailto:andrewjohn83 at outlook.com>> wrote:


Hi Dario,


Thanks for the information. I managed to get this working as long as each output is connected to a rx_streamer (i.e. QT GUI Time Sink). However, when I try to use the block within a loopback configuration (https://corvid.io/2017/04/22/stupid-rfnoc-tricks-loopback; output 0 connected to radio Tx block and output 1 connected to a sink) I get an overflow on channel 0. I am thinking that the delay in issuing the steam command for channel 0 followed by channel 1 is causing the fifo to overflow? Is it possible to enable both streams at the same time?


Andy

________________________________
From: Dario Pennisi <dario at iptronix.com>
Sent: 12 October 2017 04:54:00
To: usrp-users at lists.ettus.com; Andrew Thommesen
Subject: Re: [USRP-users] RFNoC Block with 2 inputs and 1 output

Hi Andy,
Same issue here and you can probably find a set of emails form me in this list that went unanswered. I spent quite some time debugging this just to find out that apparently this is not supported. The issue in my opinion is that sofware does not initialize the streamer for the second input unless you also have a second output. Just add a second output on the block definition and connect it to null sink and if everything else is fine it should work.
Note that since input axi bus is 64 but at 166 MHz you won't be able to have 2 inputs at 32 bit/200 MHz which is the default radios will work at so you have to change system sampling frequency to 120 MHz using the attribute in device3... Don't even try changing it in radio block as the sampling frequency there is not doing anything at all.
Good luck,

Dario Pennisi



On Wed, Oct 11, 2017 at 10:11 PM +0200, "Andrew Thommesen via USRP-users" <usrp-users at lists.ettus.com<mailto:usrp-users at lists.ettus.com>> wrote:

Hi all,

I am trying to create a custom RFNoC block with 2 inputs and 1 output. I have attached my block code as well as the XML files for the UHD and GNU radio integration. The block behaves as expected when I run the testbench. However, when synthesised and used within the GNU radio flowgraph it does not seem to output anything. Further investigation using chipscope revealed that m_axis_data_tvalid is only ever 0 or 1. In simulation, the value can be 3 when data is received on both input channels. This suggests that on the hardware I am not receiving data from both streams, which would explain why I never get any output. This may suggest that the XML files are incorrect? Does anyone know what I could be doing wrong?

Thanks,

Andy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20171014/94f50069/attachment-0002.html>


More information about the USRP-users mailing list