This doesn't necessarily mean cable reflections. could be free-space, eg. between antenna structures somehow, although that seems less likely than cables, especially given they all show same freq.
the plots are not so clear given the scale. might blow up some part of some to show the ripple better, but it is seen even at this scale. what is P-P amplitude (%), typically? is it stable in time in terms of amplitude and phase across band (ie. location of peaks in freq). if it is, bandpass calibration should help, if it isn't, then might need to calibrate BP on short timescales. or find the culprit. perhaps investigating the worst antennas would show something obvious? lastly, do you see this ripple in the antenna-based solutions of omnical? what is freq and time averaging of omnical? (each record/channel)? in fact, since it is common mode, would omnical take this out or not? where do these informal memos end up? ether? memo series? Chris > Attached is a memo from earlier this year on autocorrelations in PAPER > 128. > > ---------- Forwarded message ---------- > From: Jonathan Pober <[email protected]> > Date: Thu, Feb 5, 2015 at 3:55 PM > Subject: memo on PAPER 128 autocorrelations > To: PAPER List <[email protected]> > > > Hi everyone, > > I had a little bit of time today to do some writing, and so jotted down > some of the things we noticed looking at the PSA 128 data back in January. > There are two sections: (1) looking at a 6 MHz ripple that now shows up in > all of our autocorrelation spectra, and (2) looking into Omnical's ability > to identify and flag antennas on an initial run. I'm sending this email > to > the PAPER list, but since this is essentially the exact hardware that will > become HERA in 2 months, this may make sense as a HERA memo. I leave it > to > the powers that be to decide. > > Jonnie >
