Quick Look at first 3-port data
These webpages just give a quick rough look at the data collected with the 38GHz 3-port system during the first few days when data collection became quasi-continuous. Note that as yet the instrument is not accurately calibrated, so the following should only be used for general information purposes until more detailed and careful analysis can be carried out in due course upon the archived data. The following graphs must therefore be viewed as provisional, and treated with caution at this stage. Their main value at this point is therefore to help identify days when interesting events (or problems) occurred.
The following illustrations are low-resolution bitmap versions of vector graphs so lose some detail in order to fit easily onto a webpage.
Each illustration represents the data collected on a given day in the period from 5th March onwards. Four graphs are shown in each case. Each quantity is averaged in time-bins of approx one minute. (More precisely, each bin is taken from 4800 sample sets, taken at a sampling rate of 83.24Hz.) In each case the red line is for the left interferometer pair, and the blue line for the right interferometer pair. The graphs are labelled in the first illustration, but not in later examples so as to avoid covering the graph patterns.
The above graph is for Tuesday 5th March. On this day the transmitter was mounted on the permanent gantry and a battery was used that should have enough power to drive the transmitter for a few days. Hence this was the start of the period when we hoped to begin quasi-continuous data collection, and in particular, to collected data over night as well as during the day.
The top graph shows the nominal mean power levels. The system sensitivity and linearity is not yet precisely calibrated, so these values should be taken only for guidance, and come from assuming that I squared plus Q squared will be proportional to the power. The 0dB level assumed is an arbitrary convenient value.The top graph covers a 30dB range.
The second graph shows the rms level of power fluctuations about the mean during each time bin. Again, this graph covers a 30dB range.
The third graph shows the nominal inter-port phase difference for each port-pair. Variations in these values will be produced by any change in the apparent direction of arrival of the signal wavefront. This graph has a range of 360 deg in total.
The lower (forth) graph shows the rms level of angle variations about the mean during each time bin. This plot has a range of 40 deg. Note that the phase data is not 'stitched' so apparently large variations may occur due to cyclic errors when the mean value is near to +180 or -180 degrees. This, and other corrections, may be made upon more detailed analysis.
The above graphs are for Wednesday 6th March. The mean power (top graph) can be seen to dip a few times during the morning, and then vanishes just before mid-day. This was due to the battery voltage falling enough to allow the transmit oscillator to drift off-tune and out of our IF bandwidths. The receiver LO was re-tuned to compensate as a temporary measure, but this showed we could not rely upon a battery for obtaining 2-3 days continuous data.
The above graphs are for Thursday 7th March. The quiet period just after mid-day was when the battery supply for the transmitter was replaced with a bench supply from the mains. This took some time to fit as it was necessary to re-cable the power feed up the gantry. During this period the receiver was left on so as to obtain some zero levels. Hence the quiet period indicates receiver behaviour with the transmitter switched off entirely.
The above graphs are for Friday 8th March.
The above graphs are for Saturday 9th March. Here the weather deteriorates during the afternoon and becomes blustery with quite high wind velocities reported in weather forecasts. The high level of left-pair rms angle fluctuations in the morning should be treated with care as they may be due to unstitched phase data.
The above graphs are for Sunday 10th March. It began to snow heavily in the morning (around 6am), and by the time I woke up snow was covering the roofs and ground to a depth of between a half-inch and an inch. It continued to snow during the morning. During the period the received power faded (nominally) by at least 25dB. During the afternoon the snow ended and intermittent heavy rain began to wash it away. The dips in power in the afternoon may be associated with the rain.
Content and pages maintained by: Jim Lesurf (jcgl@st-and.ac.uk)
using HTMLEdit3 on a StrongARM powered RISCOS machine.
University of St. Andrews, St Andrews, Fife KY16 9SS, Scotland.
