discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Calling a function after every N samples


From: Michael Dickens
Subject: Re: [Discuss-gnuradio] Calling a function after every N samples
Date: Tue, 25 Jun 2019 16:13:17 -0400
User-agent: Cyrus-JMAP/3.1.6-730-g63f2c3b-fmstable-20190622v1

Hi Kevin - Thanks for the additional info. It seems that you've been doing some serious work trying out various options; very good!

If you're using gr-uhd, then you can use the data tag "tx_freq": If it's at the initial sample in the incoming buffer, then it indicates an immediate tune request at the specified tag value. If it's at the final sample, then it gets a little complicated ... you can read through this chunk of the source code here < https://github.com/gnuradio/gnuradio/blob/master/gr-uhd/lib/usrp_sink_impl.cc#L568 >. Anyway if you want to change, or the option to change, the Tx freq every N samples when using gr-uhd then one option is to create a custom block that counts the number of items going through it and every N and attaches this tag at the correct sample for what you want.

If you're not using gr-uhd, then there might still be tags such as this method that will work with the given Tx HW.

Hope this is useful! - MLD

On Thu, Jun 20, 2019, at 3:57 PM, Kevin Lee wrote:
Hello,

Thanks for the reply! More specifically, what I'm aiming for is to change frequency after transmitting some number of samples. I've tried having the sink call the set frequency commands each time 'work' is run, but that doesn't seem to help me set the length. I've tried using message passing, but that's not sufficient because that uses a sleep timer which is too inconsistent for what I'm doing. As for the vectorised data, it still doesn't time correctly since what I think the sink is doing is writing to a buffer in hardware. Though despite that, the timing result is consistent in the first method I listed. Is there a way to trigger command execution in response to specific samples? Or any way to precisely time frequency changes?

Thanks,
Kevin


reply via email to

[Prev in Thread] Current Thread [Next in Thread]