Thanks for the advice! Try to do but continuously it is showing error Param - Value(value): Value "numpy.sqrt((10.0**(-esno/10.0))/2.0)" cannot be evaluated: name 'numpy' is not defined how can i rectify this? is there any already done grc files for evaluating ccsds
On Sat, Aug 19, 2023 at 9:55 PM U L <jaredd...@gmail.com> wrote: > I'd try this: > 1. Change the random source to a vector source with vector [ 0xFF, 0xFF, > 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 ] (that's 4 x > 0xFF, 8 x 0x00) > 2. Encoder output can go to the map input, then map to a charToFloat. > CharToFloat to decoder. > 3. Remove the entire noise branch. It doesn't make sense to add noise > before the map anyways. Noise would normally be added as a float after > charToFloat. > 4. Vector source to a delay block, delay to one input of BER block. > Decoder output to pack8 to BER block. > 5. You'll have to experiment with the delay value. It's probably between 2 > and 8 bytes. > 6. If you want to try to observe the delay between the source and the > decoder, make a 2 input time sink and connect the unpack8 output after the > vector source to one time sink input, and the decoder output to the other > sink input. You should be able to see the sample delay between each block > of inputs and compute the right delay from the difference. > > Good luck, > Jared. > > On Sat, Aug 19, 2023 at 5:13 AM Jiya Johnson <jiyajohnso...@gmail.com> > wrote: > >> Any suggestions >> >> On Sat, Aug 19, 2023, 9:15 AM Jiya Johnson <jiyajohnso...@gmail.com> >> wrote: >> >>> But still in CCSDS case i am getting -0.3 only BER and not even >>> saturating in FEC extended CC case(as per CCSDS Standard book it is >>> dropping to 0 by 5dB) >>> >>> On Sat, Aug 19, 2023 at 8:54 AM U L <jaredd...@gmail.com> wrote: >>> >>>> Hmm, this has gotten more complicated since your original post. Why? >>>> >>>> A good place to start might be the example GRC as referenced on this >>>> page https://wiki.gnuradio.org/index.php?title=FEC_Extended_Decoder. >>>> You can change the FEC definitions from LDPC to CC and add a BER block with >>>> a delay b/w the input bytes and output bytes (bypass the char to float >>>> after the decoder). Once that's working, you can replace the FEC encoder >>>> with a CCSDS one. >>>> >>>> Jared. >>>> >>>> On Fri, Aug 18, 2023 at 9:05 PM Jiya Johnson <jiyajohnso...@gmail.com> >>>> wrote: >>>> >>>>> I am getting need help pls look on the screenshot >>>>> >>>>> On Sat, Aug 19, 2023 at 8:23 AM Jiya Johnson <jiyajohnso...@gmail.com> >>>>> wrote: >>>>> >>>>>> I will try and let you know >>>>>> >>>>>> On Sat, Aug 19, 2023, 8:08 AM U L <jaredd...@gmail.com> wrote: >>>>>> >>>>>>> Probably easier to use a vector source rather than text file. Vector >>>>>>> source lets you specify the bytes exactly. So you could put in [ 0xdd, ] >>>>>>> and be assured that byte will be output. >>>>>>> >>>>>>> On Fri, Aug 18, 2023 at 8:32 PM Jiya Johnson < >>>>>>> jiyajohnso...@gmail.com> wrote: >>>>>>> >>>>>>>> What I tried is 11011101( by inputting a text file )it's hex >>>>>>>> version came in the output with (3131303131313031) like that with >>>>>>>> encoding >>>>>>>> by left shifting I got the value theoretically and decode I got dd >>>>>>>> (11011101) bt this worked over windows not in ubuntu,this output I have >>>>>>>> done using messag debug option I am not aware whether it is the >>>>>>>> correct way. >>>>>>>> >>>>>>>> On Sat, Aug 19, 2023, 7:58 AM U L <jaredd...@gmail.com> wrote: >>>>>>>> >>>>>>>>> I don't understand what this means. How do you interpret it? >>>>>>>>> Jared. >>>>>>>>> >>>>>>>>> >>>>>>>>> On Fri, Aug 18, 2023 at 8:25 PM Jiya Johnson < >>>>>>>>> jiyajohnso...@gmail.com> wrote: >>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Sat, Aug 19, 2023, 7:54 AM Jiya Johnson < >>>>>>>>>> jiyajohnso...@gmail.com> wrote: >>>>>>>>>> >>>>>>>>>>> I have tried to do with a simple input 11011101 without noise >>>>>>>>>>> and with noise noise also >>>>>>>>>>> >>>>>>>>>>> On Sat, Aug 19, 2023, 7:52 AM U L <jaredd...@gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>>> Some ideas: >>>>>>>>>>>> >>>>>>>>>>>> 1. Remove the noise completely. Also use a fixed, repeating >>>>>>>>>>>> input vector rather than random bytes. Look at the actual error >>>>>>>>>>>> patterns >>>>>>>>>>>> between the input and output bits. Sometimes error patterns can >>>>>>>>>>>> clue you in. >>>>>>>>>>>> 2. Check symbol polarity (i.e. 0=>-1, 1=>1, or 0=>1, 1=>-1) >>>>>>>>>>>> 3. Try the more general (non-CCSDS) extended encoder with the >>>>>>>>>>>> same params as the decoder. >>>>>>>>>>>> 4. Esp if 3 works, compare the output of CCSDS with general >>>>>>>>>>>> extended encoder. >>>>>>>>>>>> >>>>>>>>>>>> Good luck. >>>>>>>>>>>> Jared. >>>>>>>>>>>> >>>>>>>>>>>> On Fri, Aug 18, 2023 at 8:04 PM Jiya Johnson < >>>>>>>>>>>> jiyajohnso...@gmail.com> wrote: >>>>>>>>>>>> >>>>>>>>>>>>> Hi, >>>>>>>>>>>>> Based on the given modifications still I am getting a -0.300 >>>>>>>>>>>>> range of BER constantly from 0 to 15 dB range of SNR values.No >>>>>>>>>>>>> variations >>>>>>>>>>>>> are happening. >>>>>>>>>>>>> >>>>>>>>>>>>> How can I check the functionality of the block(encoder and >>>>>>>>>>>>> decoder to check whether it's encoding and decoding properly)with >>>>>>>>>>>>> a sample >>>>>>>>>>>>> input. >>>>>>>>>>>>> >>>>>>>>>>>>> On Sun, Aug 13, 2023, 8:11 PM U L <jaredd...@gmail.com> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>>> One other thing I noticed is that the encode CCSDS, decode, >>>>>>>>>>>>>> CCSDS, and BER blocks all operate on packed bytes. (See e.g. >>>>>>>>>>>>>> https://wiki.gnuradio.org/index.php?title=Encode_CCSDS_27). >>>>>>>>>>>>>> To do this you would change your random source to output from 0 >>>>>>>>>>>>>> to 256 byte >>>>>>>>>>>>>> values. Also, considering the output of the decoder block are >>>>>>>>>>>>>> packed bytes >>>>>>>>>>>>>> you should probably skip the byte->float and binary slicer after >>>>>>>>>>>>>> your >>>>>>>>>>>>>> decoder and just input it's output directly into the BER block. >>>>>>>>>>>>>> Finally, >>>>>>>>>>>>>> looking at the >>>>>>>>>>>>>> https://wiki.gnuradio.org/index.php?title=Decode_CCSDS_27 >>>>>>>>>>>>>> page it seems the delay out of the decoder is 4 bytes, so you >>>>>>>>>>>>>> could add a >>>>>>>>>>>>>> delay of 4 between your source and BER block to align the two >>>>>>>>>>>>>> streams. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Jared. >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Sat, Aug 12, 2023 at 9:58 PM U L <jaredd...@gmail.com> >>>>>>>>>>>>>> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>>> One thing I think might be an issue is that you have no >>>>>>>>>>>>>>> delay between your source reference and your BER block. Usually >>>>>>>>>>>>>>> the FEC >>>>>>>>>>>>>>> decoders (and possibly the encoder?) may output blocks of 0s >>>>>>>>>>>>>>> before they >>>>>>>>>>>>>>> output the samples that correspond to your inputs. I don't know >>>>>>>>>>>>>>> much about >>>>>>>>>>>>>>> those FEC blocks specifically, but maybe the docs have some >>>>>>>>>>>>>>> info on their >>>>>>>>>>>>>>> delays. You can just insert a delay block b/w your random >>>>>>>>>>>>>>> source and the >>>>>>>>>>>>>>> ber block to align the input and output bits. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> Hope that helps, >>>>>>>>>>>>>>> Jared. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> On Sat, Aug 12, 2023 at 4:28 AM Jiya Johnson < >>>>>>>>>>>>>>> jiyajohnso...@gmail.com> wrote: >>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> [image: image.png] >>>>>>>>>>>>>>>> Dear community >>>>>>>>>>>>>>>> Need help for BER plot using the above flowgraph its >>>>>>>>>>>>>>>> always coming 10^-1.17 and whenever I tried to change the >>>>>>>>>>>>>>>> noise voltage the >>>>>>>>>>>>>>>> values are not changing. >>>>>>>>>>>>>>>> Please find the attached GRC file for your reference. >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> *Regards, JIYA JOHNSON* >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>
BER_CCSDS.pdf
Description: Adobe PDF document
