Rapid Prototype Array Adventures - Spring 2002

May 29 : Reports

HI - Trevor
HI - Dan
HI & Interferometry - Robert
RPA Baselines - Robert
Robert's report (.tex)
Robert's baselines plot (.ps)

May 22 : IDL Code - RQuimby


fitbaselines.pro - Grid search for best baseline and phase.
fitfringe.pro - Display data and model fringe with arbitrary phase.
getbaselines.pro - Get baseline info.
getsource.pro - Get source info.
loadfringe.pro - Load fringe data.

May 6-9 : Interferometry Daze


We set up for 4 antenna, 6 baseline observing with a set of 3-way splitters at antenna/cpu bulkhead. Baseline check will be done with Crab long track on May7.

May 2 : Interferometry Planning


A prime target for interfering next week with goal of making an image is the Cyg-A/Cyg-X region of the sky as seen at 8.5 GHz with 20m antenna at Green Bank by Langston and colleagues

. Paul D has posted his start on analysis.

April 25: Interferometry Analysis


The most authoritative writeup of how to understand this data is 2nd edition of Thompson, Moran & Swenson text; 1st edition ok too, but my page refs are to 2nd.

-ch 4: pp 86-94 give intro to geometry of baselines.
-ch 6: pp 168-175 follow signal through heterodyne receiver defining what happens to relative phase and delay.
-ch 12: pp 467-475 describe process of fitting for positions and baselines (astrometry & geodesy, resp).

Suggested approaches to get feel for data:

-collect all for given source/baseline in single file

-collect baseline data from earlier email

-understand conversion of time to UTC to LST

-fit contiguous data chunks to (a) a parabolic "amplitude term" across the 3 lags with center of the parabola defining the relative delay at that time for that baseline multiplied by (b) "phase terms" of [+sine(2.pi.rate.time+phase), +cosine(2.pi.rate.time+phase), -sine(2.pi.rate.time+phase)] temporal terms and (c) a additive constant ("DC level". This reduces all data to:
time_0 = (

April 24: I-Day


Paul and Don visit site to fix up code and launch I-Day. Baselines 2-4, 4-6 and 6-2 used with splitting now at bulkhead in control room. All's well by ~11 AM. Lots of fringes obtained during remote observing from 544 Campbell. The following show fringe record for the 3 baselines on various sources.


09:00 Cyg A/Cyg X probably shows beat between sources.
13:00 Crab 950 Jy point source.
14:00 M31 Good detection of ~150 Jy source; will be resolved at RPA
21:00 Virgo Another ~100 Jy source.

April 22: 3-Element Interferometry at RPA


Two teams visit the RPA and try out 3-element interferometry. One polarization was split at antenna base and sent down both RF lines to control room. There ant 2-4 and 4-6 and 6-2 signal pairs were sent to cpus 2,4,6. Paul created "xcorr" program to sample data as in H spectrum activity, but then produce cross correlations at, nominally, 0,+/-1 lags. Trev proceeded to install autocorrelations to provide Tsys scale. A hilarious hot/cold measurement was done duck taping the eccosorb panel to handy pitch fork, courtesy of the horse barn the RPA lab resides in. Day ground to a halt as we sorted out exact quantization scheme of A/D and c-code.

March 20: H-Day


March along the Galactic plane taking spectra
Paul's analysis with gorgeous l-v map and extraction of tangent velocities in first quadrant.

March 15: Paul, Jang, Don


Develop procedures for spectroscopy of H along galactic plane
L=160 deg; LO1=1221.4 MHz full 20 MHz spectrum
L=160 deg; LO1=1220.4 MHz full 20 MHz spectrum
L=160 deg; LO1=1221.4 MHz partial spectrum
L=160 deg; LO1=1220.4 MHz partial spectrum
L=160 deg ratio spectrum


[Spectra were scp'd from rpa0 to astro account "ay203"; check with dcb for pwd.]

The spectra are dominated by receiver gain vs frequency, G(nu). The ratio spectrum removes most of G(nu). The residual changes can be removed by baseline fit of low order polynomial with sampling below, in middle and above two lines. This is then subtracted. Then the spectrum is split, shifted by 1.0 MHz (the amount of LO shift), and differenced to produce final spectrum.

This calibration of Gain is then followed by transfer of level with hot/cold and cal on/off readings which indicate the system temperature on Kelvin scale. In ratio spectrum the system temperature is 1.00, which is subtracted via polynomial baseline fit.

Next, we need to calibrate the FFT "bin number" to frequency via, we think, 20.0 MHz/4096 although my recovery of spectrum indicates this may be off a bit -- need to check on Monday with test tone injection. Then add LO2, 192.5 MHz, and then LO1, which was either 1220.4 or 1221.4 MHz in our frequency switched observation.


And, calibration is not over! Next we need to subtract 1420.4054 MHz, the rest frequency of Hydrogen hyperfine transition, and take ratio with same and convert to "lab" velocity scale. And finally, finally, we need to determine the motion of the telescope in direction of source with respect to the "Local Standard of Rest, LSR". The Solar System moves with about 13.4 km/s wrt LSR. Only after all of these steps can we make an astronomical calculation on our spectrum. [We can use "idoppler.pro" from UG Astro Lab for this calculation.]

March 13: Paul, Alex, Don


Measure hot/cold load with cal on/off
Try out spectroscopy - problems with trigger