Hi Henry: Thanks for your reply. I am not sure what's your mean that 2^11 real FFT gives 2^10 complex channels. Could you please give me deeper explanation?
My understanding is only half output sequence is positive frequency and the output sequence is separated into two. So the pulse period is 2048/2/2/=512. By the way, does the FFT core run in real time? I mean there is no input data loss in the FFT core. If it is real time core, how is this implemented? Thanks Wan -----Original Message----- From: Henry Chen [mailto:[email protected]] Sent: Friday, 7 November 2008 9:35 AM To: Cheng, Wan (ATNF, Marsfield) Cc: [email protected] Subject: Re: [casper] Casper library basic setup Hi Wanxiang, You're seeing periodic pulses because the output is a continuous stream; you're seeing the single pulses from multiple spectra. If you look at the pulses, they are separated by 512 clock cycles, which is the the spectrum spacing for your parameters (2^11 real FFT gives 2^10 complex channels, with 2 parallel outputs gives 2^9 clocks per spectrum). --Henry [email protected] wrote: > Hi Henry: > > I resend my simple design file to you and could you please help to check > anything wrong with it? I think there are only three type of input signal: > Sync, shift and AD output. Now I can not see anything wrong with them. But > output still is period pulse instead of single pulse spectrum. > > Thanks > > Wan > > -----Original Message----- > From: Henry Chen [mailto:[email protected]] > Sent: Thursday, 6 November 2008 5:06 PM > To: Cheng, Wan (ATNF, Marsfield) > Cc: [email protected]; [email protected] > Subject: Re: [casper] Casper library basic setup > > Hi Wanxiang, > > If the sync you're setting is the simulation input to the ADC block, > then it's treated as the reset to your sync generator, not the sync > itself. In this case, when you set it to 1 you're just always holding > down the reset to your counter, so you're actually never generating > your sync pulse. In this case, the internal counters of the DSP blocks > just run on their own, so you still get output. > > Regarding the output, what do you find wrong about it? Can you be more > specific? A suggestion: look at the sync pulse output from the FFT > rather than the input. Note that the power block you're using has a > latency of 5 (3 in the multiplier, 2 in the adder), so you need to > realign the sync pulse with the power before looking at it in a Scope. > > Thanks, > Henry > > [email protected] wrote: >> Hi Andrew: >> >> Another strange thing is I still get the output no matter I set Sync to >> "0" or "1". I think Sync could be considered as a sync reset signal. I >> should not get anything out if I fix the sync to "1". >> >> Wan >> >> ------------------------------------------------------------------------ >> *From:* Andrew Martens [mailto:[email protected]] >> *Sent:* Wednesday, 5 November 2008 7:56 PM >> *To:* Cheng, Wan (ATNF, Marsfield) >> *Cc:* [email protected] >> *Subject:* Re: [casper] Casper library basic setup >> >> Hi Wan >> >> Also remember that you have a PFB so (from the memo) you should remember >> to include the number of taps in your calculation. >> >> e.g 2 order-2 reorder blocks, order-9 unscrambler, 4 input, 2048 >> channels, 2 tap PFB >> >> min sync = 2x(2x9x2048/4) = 9216x2 >> >> Good luck >> Andrew >> >> 2008/11/5 <[email protected]> >> >> Hi Andrew: >> >> Thanks for your help. According the sync_memo: >> >> My example is: 2 order-2 reorder blocks, order-9 unscrambler, 4 >> input, 2048 channels. >> >> minimum sync period = 2*9*2048/4 = 9216. >> >> And I set the counter to 9215. I guess the sync period is all right. >> >> Wan >> >> ------------------------------------------------------------------------ >> *From:* Andrew Martens [mailto:[email protected] >> <mailto:[email protected]>] >> *Sent:* Tuesday, 4 November 2008 6:07 PM >> *To:* Cheng, Wan (ATNF, Marsfield) >> *Cc:* [email protected] <mailto:[email protected]> >> *Subject:* Re: [casper] Casper library basic setup >> >> Hi Wan >> >> I had a look at your design and noticed the following; >> >> * The output of the FFT is complex data. You need to convert this >> into real and imaginary parts if it is to make sense. Use a c_to_ri >> from the Misc section of the library. >> * You need to put "Gateway Out" blocks on all signals from Xilinx >> blocks that go to 'normal' Simulink blocks. This means that all >> inputs to Scope blocks from System Generator blocks need to go >> through a "Gateway Out" block. >> * Please also re-read the memo Henry suggested on the use of syncs >> in the CASPER library. Your sync period is too small relative to the >> size of your FFT and PFB. >> * Also note that the output sync from the FFT will show you where >> your spectrum begins so you may want to look at that with the scope >> you use to look at the FFT data too. >> >> Hope this helps. >> >> Cheers >> Andrew >> >> >> 2008/11/4 <[email protected]> >> >> Hi all: >> >> I build a very basic module with PFB and FFT. But the output >> looks not right. Anybody has anyidea? What's wrong with it? >> The attached is the mdl file. >> >> Thanks. >> >> Wan >> >> >>
