From bec@haystack.mit.edu Fri Jul 16 00:53 MES 1999 Received: from dopey.haystack.edu by kuestner.geod.uni-bonn.de with SMTP (1.37.109.6/16.2) id AA22489; Fri, 16 Jul 99 00:53:49 +0200 Return-Path: Received: (from bec@localhost) by dopey.haystack.edu (8.8.6 (sendmail_886_v2)/8.8.6) id SAA13209; Thu, 15 Jul 1999 18:55:49 -0400 (EDT) From: Brian Corey Message-Id: <199907152255.SAA13209@dopey.haystack.edu> Subject: Matera X-band spurs To: vlbi@asi.it, garramone@asi.it Date: Thu, 15 Jul 1999 18:55:48 EDT Cc: colucci@mlrohp1.atsc.allied.com, petrov@kuestner.geod.uni-bonn.de, weh@gemini.gsfc.nasa.gov (Ed Himwich), dbs@gemini.gsfc.nasa.gov (Dave Shaffer), aen@dopey.haystack.edu (Arthur Niell), dgg@leo.gsfc.nasa.gov (David Gordon), tme@cygx3.usno.navy.mil (Marshall Eubanks) X-Mailer: Elm [revision: 109.18] Status: R Dear Luciano and Eus, > In order to avoid interference in X-BAND, caused by S-BAND interference > coupling in the down converter, we installed a different Down Converter for > the S-BAND only (using the same L.O. of the X-BAND D.C.). > The system works fine and we can observe no interference on the spectrum > analyzer; so we plan to perform the next CORE-67 experiment with this > arrangement. I see no problem with running the next CORE experiment with this new configuration. In fact, it may well fix the problem, but perhaps for reasons different from what you may think. I believe the X-band spurs at Matera are actually weak S-band phase cal signals leaking from the S-band IF into the X-band IF. The coupling of RFI from IF2 into IF1 is not by itself a problem. But if RFI is being coupled into IF1, then certainly phase cal will be as well, and that *is* a problem. The coupling factor of ~30 dB that you reported is just about the correct level to explain the observed spurious signals. The nearly conclusive evidence that the spurs are actually weak S-band phase cal is provided by plots of phase cal amplitude vs. phase. In the case of a spurious signal whose amplitude and phase is not changing, one would see a sinusoidal variation in amplitude with phase, with one cycle of variation over 360 degrees of phase. If you look carefully at some of the plots from the Europe experiments, you can see that one cycle actually takes more than 360 degrees. In other plots, there appear to be multiple curves on top of each other (e.g., CA063). If you "unwrap" the phase (i.e., when the phase goes through a 360-degree ambiguity, add or subtract 360 degrees to the phase to keep track of the total phase) in the latter cases, however, and then plot amplitude vs. phase, you can see clearly that the distance between peaks in the sinusoid is ~500 degrees. When this is done for CA063, for instance, there are 2 nearly complete sinusoidal cycles that span 1000 degrees. If in fact S-band leakage is the problem, one would expect the distance between peaks to be 480 degrees. [This number is calculated assuming that (1) the pcal phase changes with time due to LO phase variations, and (2) X LO phase = 4 times S LO phase.] The close agreement between 480 and ~500 degrees, and your report of 30 dB coupling between the two IF's, point strongly to S-band contamination as the problem. The channels that are affected appear to be those around 8400-8600 MHz, or 300-500 MHz in the IF. In wideband experiments (e.g., CORE-A), those channels are X3 and X4, which are the channels where Leonid found spurs. In narrowband experiments (e.g., Europe-46), Leonid found spurs in X5-X8, which again are in that frequency range. So the problem is specific not to a particular set of VCs, but to a frequency range. I would guess that that range is most affected because the IF level there is much lower than at lower X-band frequencies, so the effect of a coupled S-band signal is stronger. At frequencies above 500 MHz or below 200 MHz, there should be no S-band power, and X-band will be clean. A couple equipment questions: 1) Why is there (or at least, why does there appear to be) a 20-dB rolloff in the X-band IF level above ~300 MHz? 2) Do you have any idea what causes the 30 dB coupling between the IF's? That's a serious problem, which we need to check for at other stations -- especially at stations with receivers similar to yours. (I assume you have been using the receiver built by AlliedSignal.) There is a simple test you could do sometime to prove whether IF coupling is the problem, but you would have to restore the downconverters to their original configuration. If you do that, and if you disconnect the cable from the phase cal antenna unit to the X-band directional coupler, then normally the phase cal observed at 10 kHz in an X-band baseband USB signal should disappear. But if there is coupling, it won't. Instead, at IF frequencies of 300-500 MHz, the level will drop by only 20-30 dB. Best regards, Brian