P2160 crab single pulses
Project p2160 is looking at giant pulses from the crab
pulsar with high time resolution using a fancy oscilloscope. On previous
days an oscillation was seen in the total power output while tracking
the source (which includes the nebula). On 11mar06 cband high was
used to look at the crab. While the crab was tracked, a separate total
power time series was taken with a 500 Mhz bandwidth centered at 7000 Mhz
and a 1 millisecond time constant to check out the oscillations. The plots
show the results:
power versus time (.ps) (.pdf):
of pulses greater than 1% of Tsys: This lists all of the pulses
from the bottom plot that exceeded .01 Tsys. The columns are : hh:mm:ss.ssss
utc time, PolA total power, polB total power. If there were contiguous
time samples above .01 Tsys then only the first was listed (since the time
constant could cause many samples to be included). PolB had many more entries
than polA. When the pulsar became active, TsysA was higher than TsysB do
to the nebulas polarized power. This may have affected the .01 threshold
for inclusion in the list.
Top plot: total power versus time. The data has been divided by the median
value for the entire 2.5 hour span. The Black (pola)
and red (polB) traces
are a peak hold every 1 second. The green
(polA) and blue
(polB) traces are the median value taken once
a second. The vertical scale is in Tsys units. The variation in power with
the two polarizations comes from the polarized power coming from the nebula.
The narrow spikes are (probably) the pulses from the crab. This data has
been detected with a 1 millisecond time constant with no dedispersion.
Bottom: The baseline was removed in 1 second sections
with a 2nd order polynomial fit and then normalized to the average
power over the 1 second. The maximum value within each second was then
plotted. The spikes show the crab pulses. The black trace is polA and the
red trace in polB.
To look at the change in the total power oscillations,
the spectrum of the total power was computed in 100 second sections (we
ended up with 90 100 second sections). For each 100 seconds of data the
following was done:
The maximum value of the oscillation was about 1% of Tsys. If we were looking
at a point source, then a pointing error could put the point source on
the edge of the beam. Any small oscillations could then cause a large change
in the power from the source. Since the nebula is extended, it must depend
on the gradient across the nebula. Vertical oscillations could also move
us in focus.
The 100 seconds of data was normalized to the median value.
A 1st order polynomial was removed.
A hanning window was applied and then the data was transformed : abs(fft(tpAN)).
An image was made of the oscillation frequency versus hour angle.
The crab is at 21 Dec so the za goes down to about
3 degrees before the azimuth swings around (you see this where the total
power switches from polA to polB in the first set of plots.
frequency versus hour angle (0 to 10Hz) (.gif): This shows 0 to 10
hz versus hour angle. There is a high frequency oscillation around 5.8
hz when the dome is at high za. It moves toward lower frequency as the
telescope approaches transit and the reappears on the other side of transit.
If the oscillation is in azimuth, then the amplitude of the oscillation
should decrease as you move towards za=0 since on the sky, the displacement
goes at Ampl*sin(za). But this does not change the frequency.
frequency versus hour angle (0 to 2 Hz) (.gif): This blows up
the horizontal scale to 0 to 2 hz. The .35 Hz and .55 Hz oscillation frequencies
have previously been measured an motion oscillations during the hurricanes
info). The .35 hz oscillation remains fixed. The .55 hz oscillation
changes frequency with hour angle.
See if the amplitude of the oscillation frequency correlates with the motion
of the tiedowns.
Track a continuum point source with a pointing offset and see what the
oscillations look like. This will remove the uncertainty in the variation
of the nebulas power.