The Figures show the results of the measurements on 28jun01.
The solid line is source rising (SW portion of the dish not yet adjusted). The dotted line is source setting (SE portion of the dish which has been adjusted).
Solid lines are from 28jun01 while the dashed lines are the same source done on 02feb01.
There is a constant offset of 30 % plus an azimuth/za dependent part. In the plot versus za the dotted lines cover the portion of the dish that have been adjusted. They are consistently better (above) than the data from the unadjusted (SW) portion of the dish.
This is the size of a source needed to equal the system temperature. If dxxx is the deflection caused by xxx on the telescope output terminals then the sefd is FluxSrc*(dTsys/dSource). It depends on the source flux and the ratio of the Tsys deflection to the source deflection. There is no calValue dependence. Larger values are worse. The dashed lines versus azimuth plot the sefd measured on 02feb01.
The ratio is plotted feb/jun so that the larger numbers mean things have improved (as in fig 2.). The plot versus za shows the dotted lines (SE part of the dish adjusted) has improved more than the solid (SW).
Using dxxx as the measured telescope deflection and Cal as CalKelvins/dCal the various ratios can be written as: GainJun/GainFeb= (dSrcJun*CalJun)/ (dSrcFeb*CalFeb) SefdFeb/SefdJun = (dTsysFeb/dSrcFeb)/(dTsysJun/dSrcJun) ratioGain/ratioSefd = (dSrcJun*CalJun*dSrcFeb*dTsysJun)/(dSrcFeb*CalFeb*dTsysFeb*dSrcJun) ratioGain/ratioSefd = (CalJun*dTsysJun)/(CalFeb*dTsysFeb)The data shows that there has been a 30% increase in gain independent of the SE/SW area of the dish. There was another cycle of main tieback cable adjustments after 02feb01 and before the panel adjustments started. This could be the source of the 30% increase (although i'm a bit suspicious).
This value is plotted versus azimuth. It shows that the product of Tsys*cal increased by 8% from feb01 to jun01. This could be from the actual system temperature increasing or the cal diode output decreasing by 8%. Another possibility could be that the sidelobes in jun01 are different from those in feb01 and the fitting routine is getting a different baseline for the system temperature.
The ratio of gain improvements shows that the unadjusted panels increased 1.3 for 4500 Mhz and the adjusted panels saw a 1.40 improvement. For 5400 Mhz these numbers were 1.4 and 1.6. The higher frequency should increase faster as the surface improves (see the ruze curves.).
When quoting improvements between rise and set you should remember that the beam is over 700 feet so the illuminated area is never seeing "all adjusted panels" or "all unadjusted panels".
The highest gains are 5 K/Jy at 4500 Mhz and 3.5 K/Jy at 5400 Mhz. If you doubt the cal values, then the best SEFD values are 7Jy/Tsys at 4500 Mhz and 11 Jy/Tsys at 5400 Mhz. These values are nothing to write home about. The gain in the SE quadrant is still lower than that in the SW. This is from the pitch, roll, focus errors which are the major source of gain loss. When the pitch roll and focus corrections are made the gain increases above 7 K/Jy at 4860Mhz and the SEFD falls below 5 Tsys/Jy. (see cband prf after SE quadrant adjusted).
The 8% cal*Tsys increase feb to jun could be the cal or the actual system temperature. If it is Tsys then the gain should be ok and the SEFD will have increased. If it is the cal output decreasing then the measured gain is higher than it should be and the SEFD is ok.
plotting : x101/010628/plotit.pro