Power levels for the alfa post amps.
The power levels needed in the
post amp system are computed using a Tsys, bandwidth, and fiber optic xmter
noise temperature. The values used are:
|fiber optic xmter noise floor (fiber
Using -198.6 dBm/hz (10*log10(1.38e-23)
- 30(mW)) gives the following:
|1. input to the dewar
|2. input to post Amps (ignore losses)
|3. noise floor fiber optic xmter over 300 Mhz.
|4. Signal power needed to be 20 db above the fiber optic noise floor
|5. gain needed in post amps
The 33.6 db gain is the minimun gain so that the fiber
optics does not add to the system temperature. The maxium gain allowed
is determined by how strong the rfi is in the band. The fiber optic xmter
starts to have trouble around 0 dbm. This will give us 25 db dynamic range.
At the input to the dewar: 30 K Tsys and 300 Mhz bandwidth gives:
-198.6+ 15(30K) + 85(300Mhz)=-98.6 dBm .
With a dewar gain of 40 dB and ignoring the losses in any filters and cables.
anacom fiber optic transmitter: ac231-2.5. -130Dbm/hz + 85 db (300Mhz)=-45
dBm. This is the closest part i could find to the AC321T-CA part that we
are using. There spec sheet has -130 dBm/Hz for the output noise floor.
This is so the fiber optics contributes less than 1% to Tsys.
This is the total gain. It includes the amplifier gain and any mixer,
A table of gain,noise figure by component for the
rf processor of alfa is listed below in the table:
The first 3 columns are the noise figure, equivalent
temperature, and gain of each device. colums 4 and 5 are the cumulative
gain and temperature at the input to the device. The final column is the
devices contribution to Tsys. The Noise figure value for the fiber is taken
from -130Dbm/Hz. Some of the vendors plots shows the noise figure
as a function of fiber optical loss. The 44db could be as low as 38dB.
|mixer- hp iam8
The maximum allowed power input to the xmter is
+10dbm. The system above has -16.9 dbm at the fiber optic input.
This would give about 27 db of headroom. Of course the 10 dbm max level
may be where it melts , not where it goes non-linear.
Power levels at lband wide dewar.
A spetrum analyz we placed at the output of the lband wide dewar
and the spectrum was measured. This dewar has a gain of 40 db. The plots
show how large a signal the postamps will have to deal with:
The plots from 18sep02 have peaks at -45 dBm using a
30Khz resolution filter. The may01 data has the 1350 Mhz faa radar going
up to -22 dbm. Adding 40 db gain to -45 dBm should be ok. Adding 40 db
to -22 dBm will saturate the system. The discrepancy probably comes from
the different resololution bandwidths used. 18sep02 used 30 Khz while may02
used 3 Mhz. This is a factor of 20db difference. The power from the radars
is on the order of 1 Mhz Bw so the 30 Khz filter width was not including
all of the radar power in a single channel. The 3 Mhz resolution bandwidth
was. The other problem is trying to measure a pulsed radar that is rotating
with a swept oscillator system. You have to wait a long time to be sure
that you sweep across the radar frequency when it is at its peak.
The rf monitoring will be through a directional
coupler before the fiber optic xmter. With a 20 db coupler, the power levels
for the monitor will be centered on -45 Dbm.
Anacom 231T-CA transmitter
We are using the anacom 231T-CA transmiter. Anacom
lists the 231W
devices.These are tx/rcv pairs. Our xmter probably corresponds to the 231W-A
which is the wide band low noise figure device (in 231W). They quote a
38 dbm noise figure with a 1 meter cable. The NF of the device depends
on the optical loss (see the anacom