sbc 3,4 high Tsys. halfband fitlers 3,4
23jun11
The system temperature for interim correlator boards
3 and 4 increased
by about 3 K sometime around 21jun11 (we're not exactly sure when the
problem first occurred). After trouble shooting we found that the +5
analog voltage for halfband filters 3 and 4 (in the same chassis) had a
2V repetitive dropout. We replaced the power supply and the system
started to work.
The halfband configuration is:
- halfband filters hb1->sbc1 ,hb2->sbc2 are in the same
chassis and share the same power supplies
- halfband filters hb3->sbc3 ,hb4->sbc4 are in the same
chassis and share the same power supplies.
We took some data before and after the changes. The
setup was:
- lbw 1405 Mhz sent to all 4 interrim correlator boards with 1024
lags.
- cycle through bandwidths 25 down to 195 Khz second
- The bandwidth numbers used for the plots are:
- 2-25Mhz,3-12.5Mhz,4-6.25Mhz .. 9-.195Mhz
- For each bandwidth do a 10 second calOn, calOff observation.
- We computed the total power Tsys (using the cal value).
- We also computed the spectra of Tsys (in degK).
- Repeat this for 9 level and 3 level sampling.
- Normally PolA is about 2K higher than polB.
The data taken before
the power supply was fixed (.ps) (.pdf):
- Page 1: 9 level, total power Tsys
- The symbols used are: + polA, * polB.
- each sbc (interim cor board) is plotted in a different color.
- black,red (sbc 1,2) is about 3K less than the corresponding sbc
(3,4).
- Page 2: 9 level spectra of Tsys
- top polA, bottom polB. black,red = sbc1,2 green,red=sbc3,4
- For each sbc3,4 is higher than sbc1,2
- the noise structure is similar for both
- Page 3: 3 level, total power Tsys
- the symbols, colors are the same as page 1
- black,red (sbc 1,2) is about .5K less than the
corresponding sbc (3,4).
- Page 4: 3 level spectra of Tsys
- the colors are the same as page2.
- For each sbc3,4 is only slightly higher than sbc1,2
- the noise structure is similar for both
- Page 5,6: repeated 9 level measurements of page1,2
The data taken
after the power supplied was repaired (.ps) (.pdf):
- Page 1: 9 level, total power Tsys
- PolA is the same for all 4. sbc1 polB is about .2K less than
the other 3.
- Page 2: 9 levele tsys spectra.
- all overlay and the noise is the same
- Page 3: 3 level, total power Tsys
- Page 4: 3 level tsys spectra.
The data
taken after the power supplied was repaired repeated (.ps) (.pdf):
- We repeated the measurements (with the power supply fixed) and
got the same results.
- The polA 9level total power offset was no longer there.
Summary:
- the +5 analog supply for halfband filters 3,4 failed sometime
around 21jun11.
- this caused the Tsys for sbc3,4 to increase by about 3 deg K for
9 level sampling.
- 3level sampling has a tsys increase of about .5 K
- The noise seemed to stays the same in the four bands.
processing: x101/110622/chkicorr.pro
up_to_interim cor page
home_~phil