My block doesnt propagate tags, despite propagation policy set

[Nikoloz Glonti]

Dear Gnuradio community,

I have a problem - my block doesn't propagate tags from input 0 to 
output 0 despite i set tag propagation policy in constructor to 
TPP_ONE_TO_ONE.


Also, It can't detect tags in incoming samples when i call


get_tags_in_window(d_work_tags, 0,0,noutput_items);

for (const auto& tag: d_work_tags){
    std::string keyName = pmt::symbol_to_string(tag.key);
    if(keyName == "FirstSyncSeqDetected"){
  uint64_t aaa = nitems_written(0);
    add_item_tag(0, aaa, pmt::intern("FirstSyncSeqDetected"),
 pmt::from_double(0), d_src_id);
    }
  }



I wrote an OOT block - where I'm making moving average operation - 
simple operations in for loops .


Thank You in advance!

Nikoloz Glonti

Is there any way to tell scheduler, that our block needs certain amount of items on input?

[Nikoloz Glonti]

Dear Gnuradio community,

Is there ability to tell scheduler to call work functions only when 
ninput_items will have value I set manually in constructor?


thanks in advance!

Niko

Re: Is there any way to tell scheduler, that our block needs certain amount of items on input?

[GNU Radio, the Free & Open-Source Toolkit for Software Radio]

Hey,
Yes, use that function:
https://www.gnuradio.org/doc/doxygen/classgr_1_1block.html#a63d67fd758b70c6f2d7b7d4edcec53b3

pt., 5 sie 2022 o 13:44 Nikoloz Glonti  napisał(a):

> Dear Gnuradio community,
>
> Is there ability to tell scheduler to call work functions only when
> ninput_items will have value I set manually in constructor?
>
> thanks in advance!
>
> Niko
>
>
>

Time Partitioned FFT

[Elmore's]

Is it possible to use GNU radio to perform time partitioning of FFTs?

I would like to be able to start an FFT and run it for 160 ms. A second one 
would start 40 ms after the start of the first and run for 160 ms. A third 
would start 40 ms after the start of the second. This process would continue 
for any number of FFTs.

I hope I have correctly asked the question. I’m not even sure how you would 
limit the running time of the FFT.

--
This email has been checked for viruses by AVG.
https://www.avg.com

Re: Ubuntu 20.04 cannot find the Hackrf board?

[Cinaed Simson]

And you'll need to

  apt install gr-osmosdr

to use the hackrf with gnuradio.

-- Cinaed


On 8/3/22 23:21, Cinaed Simson wrote:

Hi George - the hackrf runs under ubuntu.

The latest version is

  release 2021.03.1
https://github.com/greatscottgadgets/hackrf/release
https://github.com/greatscottgadgets/hackrf/releases/download/v2021.03.1/hackrf-2021.03.1.tar.xz

The current firmware version on your hackrf is 2014.

You could try

   apt update
   apt install hackrf libhackrf-dev

which may give you the 2018 version for the hackrf utils and 
libhackrf-dev.


But I don't know how the firmware gets updated using apt - I've never 
used apt to install the hackrf software.


The 2014 firmware may play enough well with the 2018 software - I 
don't know. It may play well enough to indicate the hackrf works on 
ubuntu but Ideally you'd like the version of firmware to match the 
software.


-- Cinaed


On 8/3/22 18:28, George Edwards wrote:

Hello Gentlemen,

I installed VirtualBox extension pack and made the USB connection in 
VirtualBox to the HackerRF board. However, in the Ubuntu Terminal 
when I execute the command hackrf_info the info that comes up is:

hackrf_info version: unknown
libhackrf version: unknown (0.5)
Found HackRF
Index: 0
Board ID Number: 2 (HackRF One)
Firmware Version: 2014.08.1 (API:1.00)
Part ID Number: 0xa000cb3c 0x006e4756

The above shows Ubuntu is having trouble reading the top two version 
information and as a result when I try to run the Gnuradio flowgraph, 
it gives a message that it fails to open HackRF.


George



On Wed, Aug 3, 2022 at 10:32 AM James Anderson  wrote:

You might want to try running hackrf_info with superuser
privileges using sudo, i.e. "sudo hackrf_info". Many/most
peripherals require this in Linux by default.

Additionally, use of a VM may limit the sample rate you can
achieve since there is some overhead passing through the USB
connection. Consider installing Linux directly on your machine.


On Aug 3, 2022, at 6:12 AM, George Edwards
 wrote:


Hello GNURadio Community,

I built a grc flowgraph in Gnuradio 3.9.5 on Ubuntu 20.04 inside
Microsoft VirtualBox. I have a HackRF One radio hardware. I
installed the hackrf drivers in Ubuntu with command: sudo
apt-get install -y hackrf
and confirmed the installation. I connected the HackRF One board
to my computer and in the Terminal prompt entered the command
hackrf_info and received the response that it does not see the
Hackrf board.

And true to form when I ran the flowgraph, I get the following
error message:
RuntimeError: no hackrf device matches

Why is Ubuntu 20.04 running inside MS VirtualBox not seeing the
HackRf board with the HackRf drivers installed. And, how do I
resolve this issue?

Will appreciate any help to resolve this issue.

George

Re: Time Partitioned FFT

[Chris Kuethe]

Why?

How would this work on a 2-core ARM SBC vs some fancy 32-core workstation?
What about the incoming sample rate? Are you getting 2Msps from an RTLSDR,
or 200Msps from a very powerful instrument?
And how many points is your FFT?
What are you doing with the output of the FFT?
1000ms / 40ms interval = up to 25 parallel FFT threads... do you have the
memory bandwidth to support this?

In other words: what problem are you trying to solve?

On Fri, Aug 5, 2022 at 6:22 PM Elmore's  wrote:

> Is it possible to use GNU radio to perform time partitioning of FFTs?
>
> I would like to be able to start an FFT and run it for 160 ms. A second
> one would start 40 ms after the start of the first and run for 160 ms. A
> third would start 40 ms after the start of the second. This process would
> continue for any number of FFTs.
>
> I hope I have correctly asked the question. I’m not even sure how you
> would limit the running time of the FFT.
>
>
> 
>  Virus-free.
> www.avg.com
> 
> <#m_-4793840578475885421_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>


-- 
GDB has a 'break' feature; why doesn't it have 'fix' too?