Hello Rolando, You expect to see a DC bin, i.e. a large value in channel 0, and I have often seen a large-ish value in bin 1 as well. After that it will drop off.
If you're seeing a peak in something other than 0, there may be something wrong. Those spikes at the high end of your spectrum look a little bit fishy to me, though (can someone correct me?) this may depend on which Nyquist zone you're sampling in. Regards, James On Mon, Mar 12, 2018 at 5:00 PM, Rolando Paz <flx...@gmail.com> wrote: > Hi. > > With the help of Indrajit, I managed to understand where to place some > delays to adjust the spectra position... > > My test tone is 60MHz. > > I still see a signal to the left of the graph, does that mean that I still > have to adjust the simulink design more? > > Regards > > Rolando > > 2018-03-10 19:51 GMT-06:00 Jack Hickish <jackhick...@gmail.com>: > >> Hi Rolando, >> >> My first suggestion would just be to carefully look at your design and >> follow the sync and data paths looking for issues. Otherwise you can >> simulate a signal in a particular FFT bin and check with a scope to make >> sure it stays where it should relative to the sync pulse. For these kinds >> of simulations, sometimes I just make the simulation input a constant >> (i.e., a DC signal). This should result in a spike in FFT channel 0, which >> occurs the clock after the sync pulse. >> >> Cheers >> Jack >> >> On Sat, 10 Mar 2018 at 17:42 Rolando Paz <flx...@gmail.com> wrote: >> >>> Very interesting. Thanks Jack. >>> >>> Is there any way to find the place and value of the latency that I must >>> remove or add, to the synchronization pulse inside my design? >>> >>> Regards >>> >>> Rolando >>> >>> >>> 2018-03-10 18:47 GMT-06:00 Jack Hickish <jackhick...@gmail.com>: >>> >>>> Hi Rolando, >>>> >>>> This sort of channel number offset issue usually indicates a >>>> misalignment between the sync pulse in the design and data where your data >>>> goes through an operation that has some latency, and this latency isn't >>>> compensated for in the sync signal. >>>> One clue is that there is usually a spike in FFT bin 0 (i.e., the DC >>>> bin). In your plots this spike appears at the end of the spectrum for the >>>> per-antenna plots, and seemingly at bin ~2 in the beamformer plot. >>>> >>>> You should fix this in your simulink design, by adding or removing >>>> latency in the sync or data signals to keep them aligned. You could just >>>> shift your spectra in software, but that's a bit of a hack -- really you >>>> should just fix the hardware bug. >>>> >>>> Cheers >>>> Jack >>>> >>>> >>>> On Sat, 10 Mar 2018 at 15:54 Rolando Paz <flx...@gmail.com> wrote: >>>> >>>>> Hi Jack >>>>> >>>>> I did some tests with my beamformer design (4 inputs). >>>>> >>>>> Currently I only have a 70MHz test tone at the A and B inputs. I do >>>>> not have anything connected at C and D inputs . >>>>> >>>>> The tone at A and B inputs is slightly offset to the left with respect >>>>> to the 70MHz signal. >>>>> >>>>> In the C and D inputs appear some signals that I do not know why they >>>>> appear. >>>>> >>>>> Do you know why the spectrum can move? >>>>> >>>>> In the case of the beamformer signal, it appears displaced to the >>>>> right of the 70MHz tone. Why does this happen? >>>>> >>>>> Is this corrected in the spectrometer design (matlab design) or is it >>>>> corrected with python? >>>>> >>>>> Best Regards >>>>> >>>>> Rolando Paz >>>>> >>>> >>> > -- > You received this message because you are subscribed to the Google Groups " > casper@lists.berkeley.edu" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to casper+unsubscr...@lists.berkeley.edu. > To post to this group, send email to casper@lists.berkeley.edu. > -- You received this message because you are subscribed to the Google Groups "casper@lists.berkeley.edu" group. To unsubscribe from this group and stop receiving emails from it, send an email to casper+unsubscr...@lists.berkeley.edu. To post to this group, send email to casper@lists.berkeley.edu.
