|
From: | Bakshi, Arjun |
Subject: | Re: [Discuss-gnuradio] X310 with 2 WBX in full duplex |
Date: | Mon, 20 Nov 2017 18:14:19 +0000 |
Hi,
Thanks for the info about the sampling rates. That was the source of the issue.
The issue was with the output rate of the DUC block, which should be set to 200MHz. Changing that to 200e6 fixes the underflow issues. I guess that the input rate is the sampling rate that I'm used to setting in the N210. I can now run 2
TX blocks without a problem.
The sampling rate in the RFNoC-radio(tx) block can be set to what you want, but it'll get reset to 200e6 along with a warning saying so. I am still having trouble with the RX blocks. -- When I add an RX flow to a graph with 2 TX flows in it, I get underflows again. The application does not hang though.
-- When I have 2 RX flows to the graph, I get the following error: self.device3.connect(self.uhd_rfnoc_streamer_radio_2.get_block_id(), 0, self.uhd_rfnoc_streamer_ddc_1.get_block_id(), 0)
File "/usr/local/lib/python2.7/dist-packages/ettus/ettus_swig.py", line 1670, in connect return _ettus_swig.device3_sptr_connect(self, *args) RuntimeError: RuntimeError: On node 0/DDC_0, input port 0 is already connected. terminate called after throwing an instance of 'uhd::assertion_error' what(): AssertionError: [0/DDC_0] Attempting to disconnect input port 0, which is not registered as connected! It seems to have an issue with the lower DDC block. I changed the setup to have only 1 DDC block with 2 channels, but that resulted in an overrun error like "Ooverrun on chan 0".
How can I fix these issues?
Image of error prompt, and grc file attached.
Thanks,
AB
From: Michael Carosino <address@hidden>
Sent: Monday, November 20, 2017 12:31:52 AM To: Bakshi, Arjun Cc: address@hidden Subject: Re: [Discuss-gnuradio] X310 with 2 WBX in full duplex Hi,
looks like gnuradio is freezing due to the massive amount of underflows ocurring. First try replacing the constant source with two separate analog fast noise sources connected to the dma fifo - this might help if indeed the problem is that the PC is being
too slow to generate samples. Secondly, revert the dma fifo depths to the default 32M and see if it has any effect but I suspect the first option should fix your issues. Also it's definitely helpful to test each single tx chain on its own to make sure it works
as expected before trying both together!
Assuming you are using an x310, the sampling rate on the radio block should be set to 200 megasamples/per sec, similarly the DUC output rate should also be 200 megasamples/per sec. On the receive side, the radio block's sampling rate should again be set
to 200megasamp/sec with the DDC input rate being set to the same as well. Also the analog bandwidth in the context of the tx side just adjusts some filters I believe - so I'd leave it at the default of 56Mhz. On the rx side I'd also leave it at the default
of 56Mhz unless you have some interferer in that range that you'd want to avoid.
On Sat, Nov 18, 2017 at 4:53 PM, Bakshi, Arjun
<address@hidden> wrote:
|
2rx.png
Description: 2rx.png
test_X310.grc
Description: test_X310.grc
[Prev in Thread] | Current Thread | [Next in Thread] |