discuss-gnuradio
[Top][All Lists]
Advanced

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

Re: [Discuss-gnuradio] Timed USRP source commands and tags


From: Marcus Müller
Subject: Re: [Discuss-gnuradio] Timed USRP source commands and tags
Date: Wed, 23 Nov 2016 17:35:51 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.3.0

Dear Meelis,


On 23.11.2016 16:28, Meelis Nõmm wrote:
Hello

We are building a frequency scanner. As suggested in many posts, like in [1], one should use timed command messages.

Now my questions are:
1. Do I understand it correctly that the gnuradio USRP Source block does not add a tag when the center frequency changes?
That's correct; since the GNU Radio blocks run on the PC, they simply *can't* know when something like changing the centerfrequency happens on the USRP – because that might happen with a command time, and the command time refers to the clock on the USRP, not to the PC's clock, and these two clocks are _completely_ independent.
2. If that is correct then we just have to trust that the frequency changed at the requested time?
Yes. But that's UHD's job. You also have to trust your operating system to open a file and give you its content when you want to read a file...
So, unless a "L" (for: "L"ate, ie. a timed command arrived at the USRP *after* its command time had already passed) happens, you should be fine.
3. Does the USRP output the center frequency change information and it is lost along the line or is it never given out (just would imagine that the FPGA should send this vital information with the data packages)?
There's no such information sent by the USRP.
4. This would mean that the data flow between work functions, knows nothing of the center frequency alteration mid block (unless we attach a tag to the "correct" sample)?
Yep – but that's the beauty of the timestamps in the first sample of a reception – because you know the sampling rate, and you know the first samples' exact time from that timestamp, you can infer any subsequent sample's (USRP-clock-relative) timestamp by a simple division of the number of samples processed since that timestamp and the sampling rate!

Best regards,
Marcus

Just trying to understand the situation and our options
Thank you
Meelis


_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio


reply via email to

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