Sbc4 polA highTsys

•  The temperature was stable over many days.
• The amount of temperature difference depended on the bandwidth the interim correlator was using.

• The setup
• cband at 5Ghz was used
• the same signal was sent to sbc1,2,3,4
• 25 MHz bw, calon, caloff
  
• when switching the input cables, the power levels were set manually
• The off cal  was scaled to kelvins using the cal deflection.
• The  spectra from each sbc was ploted using a different color:
• sbc1, sbc2, sbc3, sbc4
  
• Frame 1,2: polA
• frame 1 signal goes straight through. 4A has higher Tsys
  
• frame 2: switch inputs 3a,4a to the digitisers. sbc 4A still high.
• this implies that the problem is after the digitizer
• Frame 3,4 polB
• frame 3: pol B signal goes straight through
• frame 4: pol B signal goes straight through.
  

• This shows that the problem is in the half band filter chassis.
• Cycling the power on the half band filter chassis did not fix the problem.

•  removed the half band filter chassis from the rack
• opened it up, looked at it, and then closed the chassis
• reinstalled the chassis in the rack
• After this the problem went away.

• Page 1: Tsys 9 level  bandwidths 25 MHz to 195 Khz.
• Each board is plotted in a different color
• The bandwidth numbers are 50MHz/(2^}(n-1) ,, n=2..9
• + is polA, * is polB
• Page 2: over plot the polA spectra
• Page 3: Tsys 3 level samlpling.. bandwdidths=25,....
• Page 4: over plot the 3 level spectra.
• 9 level polB sbc1 has a Tsys that is about .4 K larger than the other values ??

• 22jun11 Tsys function of bandwidth. Problem was half band filter power supply
• 15Aug11: spike in center sbc1 polB. problem went away when we removed and the re installed the chassis.

Summary:

• 18jun12: tsys sbc4A tsys increased. The increase was  a function of the cor bandwidth used.
• The problem was in the half band filter chassis.
• A higher tsys could result if the half band filter was configured incorrectly and a wider bandwidth was pushed through when a narrowing was expected.
  
• 27Aug12: removing the chassis, replacing the chassis made the problem go away.
• Cycling the power did not make the problem go away.
• on 15Aug11 we had a spike in the center of the band. It went away when the chassis was removed, re installed.
• switching the rf input cables  did not change the problem.
• One of the control cables to the half and filter might be making lousy contact. Removing and reinstalling the chassis could reseat it.
• The datataking code used is under usr/x101/src/src.27Aug12.