Experiments with profiling galactic neutral Hydrogen
For the purposes of galactic neutral hydrogen measurements the existing RF front end had to be modified
to incorporate a crystal locked 1st oscillator, because the free running VCO could slowly drift with tempe-
rature by up to 3 MHz, making measurement of the hydrogen rest frequency (1420.405MHz) impossible.
The mixer/oscillator/multiplier module I.F.output is split and amplified before going to two different I.F.
amplifiers, one for the phase switched channel, the other for hydrogen measurements.
The hydrogen channel is actually a complete , specialised I.F. receiver covering 131.8 - 143.0 MHz.
Its 2nd oscillator is slowly ramped from 131.3 - 132.3 MHz to perform the spectral scan across the
hydrogen bandwidth. The signal itself is actually quite weak within the 50kHz filter bandwidth of the
10.7MHz 2nd I.F. I decided to employ frequency shift switching in synch. with an integrating post-
detector switch to improve the S/N ratio. The offset frequency switch operates at approx. 300Hz.
Below are some PicoLog frequency scans. The vertical axis is 0 - 2.5Vdc, the horizontal axis is also
in Vdc, but is actually 1420MHz +/- 600kHz and is scaled to display this with a look-up table in X-Y mode.
Each scan is started manually by enabling Pico data collection, then I switch on the scan timer .
Scans take ~ 12 mins and are integrated at 30 secs "on the fly" . The detector is set for ~30mV output
to the back-end ccts. The scan VCO's stability is of course not as good as a PLL , but I have tested it
from cold start it drifts ~ 40kHz max. This is not too bad, because the I.F. filters are 50kHz wide.
I could make it narrower, but the actual signal gets perhaps a bit too weak.
Here are 11 scans on one page, Cygnus region, starting from just before the spiral arm gets into the
antenna beam, following it through the strong part, then leaving it again. It is in fact not very easy to
actually find a region of sky without measurable hydrogen.....
Conclusions :
The experimental set-up ,although working ok, is in reality flawed in a number of ways. The greatest
problem is that the antenna is operated in drift scan mode, no tracking possible at this stage.
Because each scan requires 12 mins to complete, the scan centre moves during this time as well.
For valid profiles the antenna needs to be tracked ! The other flaw is the 2nd oscillator, which is free-
running. Although this could be overcome by replacing this oscillator with a PLL type, there is no
incentive to do so until the antenna can be made to track.
The main objective however was achieved, which was to prove the combination of ramping and frequency
shift switching can be made to work easily, resulting in greatly increased sensitivity. The results so far
obtained indicate that even narrower bandwidths could be used, maybe ~15kHz being the limit.
