Hi Guys, Let me add my two cents here:
The chip synchronization can take care of the symbol synch since the chip correlation does not produce maximum values unless the symbol timing is correct. Normally, and what I mean by normally is typical cases to the best of my knowledge, the synchronization is done in two stages (fast and fine stage) where the fast stage perform synchronization with a step of half chip duration. Some design decision need to be made regarding the source of your timing extraction, will it be: 1- preamble (contains synch sequence at the beginning of each given number of data packets) 2- use separate synch channel (the way it is done in CDMA) 3- use blind synchronization or NDA (non-Data Aided) synch, which I doubt it is gonna be option for a "simple" transceiver. I hope that can help a little! Kind regards, Yahia Tachwali ________________________________________ From: [email protected] [[email protected]] On Behalf Of Firas A. [[email protected]] Sent: Saturday, January 10, 2009 11:02 PM To: [email protected] Subject: Re: [Discuss-gnuradio] Spread Specturm implementation in FPGA Hi, > Yong J. Chang" wrote: > Hi All, > > I've worked on Embedded system and FPGA systems for several years and I > joined the team working on GNU Radio > project. Because of my lack of > knowledge about digital communication, I need some help of you all. > > The goal is simple PAM modulated communication with M-sequence spreading. > So far as I know, this can be done in > host PC level. However, our goal is that this is implemented in FPGA level > in real-time because we have to determine > exact time of packet arriving. > > The question is following. > > If I implement the despreader ( is it called synchronizer? including > detecting and tracking ) in FPGA, should I perform > symbol timing recovery in host PC again? If we use the despreader and can > detect the peak of correlation between > received signal and spreading code, I don't think that I have to recover > symbol timing again because the despreaded > signal is already symbol-spaced signal captured at nice time. But I'm not > sure that I'm right or wrong. > > And I hope that this can be done in Cyclone FPGA!!!! > > Thank you all! > > Y. J. Chang I'm not sure how it is possible to do the spreading with PAM, but I think doing spreading with PWM is much easier. Also to conserve space in USRP FPGA, you have to use the image with 1 RX and 1 TX. This will free about 55% of FPGA resources. Best Regards, Firas -- View this message in context: http://www.nabble.com/Spread-Specturm-implementation-in-FPGA-tp21380279p21396189.html Sent from the GnuRadio mailing list archive at Nabble.com. _______________________________________________ Discuss-gnuradio mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnuradio _______________________________________________ Discuss-gnuradio mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
