youlizhao / ofdm Goto Github PK

For commit 69d84b6, "import raw" has error: undefined symbol: _Z13digital_crc32PKhm.

Des.: (1) If we use higher decimation/intepolation value (lower bandwith, e.g., decim: 100 -> 500), we can not find peak any more using STS or LTS (FTW project file); (2) Sometimes, higher bandwidth does not mean the existing of peak. Weird; (3) If we use sleep mechanism to send periodic packets, sometimes not all packets are sent periodic, e.g., we only get 4 packets with equal IPI. Maybe others are lost in queue or struck? (see gr_msg_queue.cc Line 54: while(full_p())) d_not_full.wait(guard)).
Guess: It is caused by CFO. Need validation.
Plan: Sync two usrps, and send the same file with same bandwidth to see whether peaks exist.
Task: Sync two USRP programs;

(1) Why can we find two preambles using benchmark_tx.py + benchmark_rx.py?
(2) How to understand sig[0]?
(3) Sleeping in Tx leads to lower PRR in Rx, Why?

When I install ftw_80211_ofdm_tx files, the modules are always installed in pythonX*Y/site-packages. To avoid adding new PYTHONPATH, I want to install them into dist-packages. The reason why I want to do that is because when I write from gnuradio import ftw. The program is not so smart to recogize it should use dist-packages/ or site-packages. In other words, the system only recognize one folder.

To tackle this problem, I change site-packages to dist-packages in ftw_80211_ofdm_tx installation files, e.g., configure. I can use find . | xargs grep "site-packages" to find all files. However, it does not work. After configuration, site-packages appear again. It is weired.

Any hints?

Bug Description: If I send two packets using while, I can receive multiple packets (much more than 2 packets). It seems that it is a stream.

I guess something is wrong with the repetition parameter. I should fix this bug last time :-(

Phenomenon: if I use benchmark_txrx.py to setup Tx/Rx, no problem. However, if I first run benchmark_rx.py, then benchmark_tx.py, the program will tell me that no usrp is found. It is weired, since benchmark_txrx.py still uses two paths, one sink and one source at the same time. Why?