I had a similar problem a few weeks back.  I don't remember all the details, 
but I was only able to get it to work for smaller FFT lengths.  In addition to 
keeping spp consistent, I'd also try shorter FFT lengths (with corresponding 
shorter spp).  I was working with UHD 4.4 and GNURadio's main branch.

-- Clint

On August 14, 2023 2:35:19 PM UTC, Paul Atreides <[email protected]> wrote:
>Luca:
>At one point I was trying a Fosphor FPGA image and I remember it was important 
>to set the spp block parameter to be equal in all the blocks that would accept 
>that parameter. So spp=1024 or whatever your FFT size is. That may have 
>changed since 4.0 but that was the case in the past. 
>
><end transmission>
>
>> On Aug 14, 2023, at 10:29, Bachmaier, Luca 
>> <[email protected]> wrote:
>> 
>> 
>> Hello everyone,
>>  
>> I’m currently developing a GNU Radio flowgraph with RFNoC 4.3 which uses the 
>> RFNoC FFT block. According to this workshop 
>> (https://www.youtube.com/watch?v=4XXqk0yGvCI @ 18:10) there are five runtime 
>> parameters you have to set for the FFT block: magnitude, direction, length, 
>> fft_scaling, shift_config.
>> I’m not exactly sure where in the flowgraph I can set these properties. E.g. 
>> is setting the “Block Args” parameter of the FFT block to 
>> “magnitude=complex,direction=1,length=1024” correct? If not, what is the 
>> right way to set the runtime parameters?
>>  
>> After setting them as I described I get two python errors when trying to run 
>> the flowgraph:
>>  
>> 1. Setting magnitude=complex causes this:
>>              RuntimeError: RuntimeError: Property magnitude:RuntimeError: 
>> Cannot convert `complex' to int!
>> In the flowgraph, the output of the RFNoC FFT block is connected directly to 
>> an RFNoC Rx Streamer block. It seems that all default RFNoC blocks only 
>> accept an int input. This error seems strange and that’s why I doubt that I 
>> set the magnitude parameter correctly.
>>  
>> 2. No matter what parameter I set for the FFT, I get the following error:
>>              RuntimeError: ValueError: samples per package must not be 
>> smaller than atomic item size
>> I was not able to find any sufficient information about this online 
>> unfortunately.
>>  
>> I hope I was able to describe my problems clearly and would be glad to 
>> receive help regarding any of those.
>>  
>> Thank you in advance and regards
>> Luca
>> _______________________________________________
>> USRP-users mailing list -- [email protected]
>> To unsubscribe send an email to [email protected]
_______________________________________________
USRP-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to