Lbw upstairs iflo cable check
oct06
links:
The setup
The test results
Conclusions
plots:
25
Mhz total power during the tests (.ps) (.pdf)
The setup (top)
People have been having baseline problems with lbw polB.
On 26oct06 and 27oct06 we tested the cables in the upstairs if/lo that
lbw used. We tapped on cables/chassis to see if vibrations changed the
signal strength. We also loosened cables by 1/2 of a turn to see how sensitive
the power levels were. While the cables were being tested, the interim
correlator was run with a 25 Mhz bandwidth centered at 1405 Mhz. Data was
dumped once a second. The identical data was sent to 4 subcorrelators.
lbw receiver components: (top)
The lbw receiver components up in the dome contain:
-
dewar
-
lbw pol box. This optionally converts from native linear to circular polarization.
-
lbw filter bank. There are two boxes (connected by short cables) that make
up the filter bank.
-
iflo Chassis. Standard chassis with equipment shared by all receivers.
-
Rf switches to select 1 receiver
-
transfer switch to swap polA,polB cables
-
rf attenuator
-
mixer chassis. There is a 750 Mixer chassis and a 10 Ghz upconverter mixer
chassis
-
if attenuator
-
fiber optic transmitter.
The test results
(top)
On 26oct06 we started testing the cables on the lbw pol box, lbw filterbank.
On 27oct06 we looked at the iflo rack standard chassis. The plots shows
the 25
Mhz total power during the tests (.ps) (.pdf).
The black trace is polA, the red trace is polB
lbwPolBox to lbw (page 1)
-
Top: lbwPolBox to lbwFilterbox1 The cable between the lbwPol
Box and the lbwFilter box1 was loosened 1/2 a turn. PolA power decreased
be 6%, polB did not change.
-
Bot: clean & reverse lbwPolBox->lbwFilterbox1: The pol A cable
that showed the jump in 1 was cleaned. It was then reinstalled with
the ends reversed (polBoxSide went to filterBox1 side). The polA cable
was loosened on the polBox side (dashed green..it
jumped 1%) and then on the fitlerbox1 side (dashed
blue, it jumped .4%). Cleaning and retightening seemed to improve
things. The poor connection is probably on the cable and not the polbox/filter
box since the largest response stayed with the cable end.
lbw Filterbox1 to filterbox2 (page 2)
-
Top 1360 to 1470 filter: Filter 6 was selected.
-
filter box 2 output cables loosened 1/2 turn (dash black and
red lines). No change in power
-
fbox1 to fbox2 cables, fbox1 side. Little or no change (no vertical
flags in plot)
-
Fbox1 to fbox2 cables, fbox2 side: PolA (dashed
green) dropped by 6% when loosened 1/2 turn. PolB (dashed
blue) showed little change.
-
Center 1100 to 1800 filter: filter 5 selected:
-
fbox1 side: PolA (dashed black) dropped
< .2%. PolB (dashed red) dropped by 4%.
-
fbox2 side: PolA (dashed green) dropped
< .5%. PolB (dashed blue) dropped by .5%.
-
reverse polB fbox1 to fbox2 cable: The polB cable was removed, cleaned,
and then reinserted with the ends switched. PolB
fbox1 side and polB fbox2 side dropped
by about 2% when the connectors was loosened 1/2 a turn. The poor connection
is probably in the polB cable and the connector on the fbox2 side.
-
Bottom swapped polB cable filter 3 (1230-1470) and polB filter 5 (1100-1800).
-
The polB cable from filter 3 (1230-1470) was moved to polB filter 5 (since
filter 5 is used more than filter 3). Loosening the cable on filter 5 polB
fbox1 side showed no change. Loosening polB filter 5 on the fbox 2 side
(dashed red) dropped the power about 1.5%.
This is consistent with the center plot that showed that the poor connection
was the cable and the connector on the fbox2 side.
750 mixer chassis: (page 3):
The 750 mixer chassis is in the iflo rack.
-
Tapping on 750 mixer chassis front panel: We could not take data
with the 750 mixer chassis open since the cables would not reach. Instead
we tapped on the front panel. PolB (dashed green)
increased in power by 4% after we started tapping. The chassis was open
and we found 3 cables that were loose. Tapping on the front panel must
have decreased the attenuation in the loose connector. The cables were
tightened and the chassis reinstalled. Tapping the front panel no longer
changed the signal level.
Conclusions: (top)
-
3 loose cables were found in the 750 Mhz mixer chassis. This could have
caused unstable power levels if the observation used this mixer.
-
The lbwpolBox to fbox1 polA cable showed a 6% drop in power when loosening
1/2 turn. After cleaning and reversing the cable, the drop was 1%. The
cable was probably loose or dirty.
-
filter 6 (1360-1470) fbox1 to fbox2 polA dropped by 6% when loosened 1/2
a turn. I'm not sure if we then cleaned or swapped it. Probably need to
recheck it.
-
filter 5 (1100-1800) fbox1 to fbox2 polB dropped by 4% when loosened 1/2
a turn. Swapped polB cable with filter 3 (1230 -1470). Now only drops 1%
when loosened on fbox2 side.
-
replace the cable that is in filter 3 polB (1230 to 1470)
-
replace connector fbox 2 side polB filter 5.
processing: x101/061026/lbwchk.pro
<- page_up
home_~phil