Search Engine **www.edaboard.com**

11 Threads found on edaboard.com: **Coherent Fft**

Which windowing function are you using? Also can you double check if your **coherent** sampling is correct.

Analog Integrated Circuit (IC) Design, Layout and Fabrication :: 06-10-2015 16:41 :: kemiyun :: Replies: **5** :: Views: **1017**

Hello guys,
I have implemented an ideal ADC+DAC converter and I did an **fft** analysis. I did a test with **coherent** sampling with rectangular window and everything was ok. Input freq. 1,376...MHz, 500mV, sinewave and a starting phase of -90?. That gave Me 7,96bits (8bits ideal).
Then I changed the phase to -60? and rectangular window and gave me

Analog Circuit Design :: 03-15-2013 00:41 :: AMSA84 :: Replies: **0** :: Views: **1478**

I'm trying to measure the SNDR of a SHA, but I'm having issues with my measurement setup. The sampling clock is 5MHz, and the input frequency is 2.0068359375 MHz, **fft** length 1024 for **coherent** sampling (411 cycles). I am using output strobing to ensure there's no interpolation when the **fft** samples. I'm trying to make sure my (...)

Analog Circuit Design :: 11-25-2011 20:32 :: allanvv :: Replies: **3** :: Views: **768**

I'm getting the raw captured data from a RT signal acquisition card. I have to perform time domain averaging so (imho) there is no question of I**fft**. Provision of storage is there (FIFO, RAMs) but how can I go about the algorithm?
Need to perform **coherent** integration (pulse integration) for N pulses to improve SNR by N times.

PLD, SPLD, GAL, CPLD, FPGA Design :: 04-07-2011 04:48 :: ravics :: Replies: **10** :: Views: **2557**

If you do not use "**coherent** sampling" you need to apply a window function to your signal before the **fft**, see below.nsignal=4096;
fs=44100;
f=1001*fs/nsignal; % **coherent** sampling i.e. integer number of cycles (1001) in nsignal samples
%f=1000; %non-**coherent** sampling. Use hann window below
t=/fs;
y=1*sin(2*pi*f*t);
w=ones(nsignal,

Analog Circuit Design :: 11-04-2010 20:49 :: JoannesPaulus :: Replies: **1** :: Views: **2190**

Unless you use an adequate windowing function for sampling then **coherent** sampling is needed to increase your **fft** spectral resolution.
Without windowing or **coherent** sampling you'll have waveform discontinuancies and see spectral leakage / artifacts in your **fft** data i.e You see spreading of energy for one frequency across (...)

Analog Circuit Design :: 11-01-2009 18:17 :: hage-bai :: Replies: **2** :: Views: **1893**

hi
i want to do **coherent** demodulation in matlab using **fft** at each step can anybody plz help me with that?

Digital Signal Processing :: 10-26-2008 17:03 :: icaniwill :: Replies: **1** :: Views: **1767**

Remember that the output of a Delta-Sigma is typically not periodic, even if your input is. As a consequence, you won't be able to use **coherent** sampling (sample an integer number of cycles). When you generate your **fft**, this will produce spectral leakage -- part of your signal will "spill over" into adjacent frequency bins. So by using the signal bi

Analog Integrated Circuit (IC) Design, Layout and Fabrication :: 09-17-2008 16:46 :: lladnar23 :: Replies: **3** :: Views: **1141**

If I understand ur problem correctly .. U r trying to Acquire a PN code having some doppler (frequency offset) from base band.
The best approach I can think is **fft** acquisition technique if u r using DSP processor for acquisition . U take **coherent** data and then do non **coherent** integration for reducing noise, finally perform (...)

Digital Signal Processing :: 01-21-2008 07:07 :: priyankguddu :: Replies: **4** :: Views: **1602**

what is the resolution in ur design?
if the resolution is small then how much is the quantization noise level? if big then how about the thermal noise in the sc? how about the dc gain of ur opamp?
if everything ok, then you have to increase the # of points in **fft** and meantime you have to use **coherent** sampling,that is u have to make sure fin/fsamp

Analog Circuit Design :: 03-23-2006 17:57 :: arsenal :: Replies: **3** :: Views: **1628**

Hi all,
I am design a 10bits pipeline ADC and by followingprevious articles I use Matlab to do **fft**. But here are the problems I met.
fin=100Khz, fs=10Mhz,
N=8192(number of data points in **fft**),
M=83 (integer number of cycles within the sampling window),
First I used **coherent** sampling method to do **fft** and I got a (...)

Analog Circuit Design :: 01-17-2006 09:17 :: yushen_yang :: Replies: **1** :: Views: **1545**

Previous
1
Next

Last searching phrases:

nobody | nobody | nobody | nobody | nobody | seven | seven | first course | seven | nevertheless

nobody | nobody | nobody | nobody | nobody | seven | seven | first course | seven | nevertheless