And not see

Jasper Wall served as Director of the Royal Greenwich Observatory and of the Isaac Newton Group of Telescopes, La Palma. He is now Visiting Professor, University of Oxford, and Adjunct Professor, University of British Columbia.

In 1965 Donald Chu, Allan Yen and I made extensive sky brightness measurements at 320 and 707 MHz. Comparison told us that something was wrong with the zero point, wrong by the same few degrees at each antenna and at each frequency. Here is the story.

Engineering was in my blood, via father and grandfather. I grew up in the Ottawa Valley, in a happy and stimulating household in which the mantra was "This works so well we must take it apart to see why." Clocks, toasters, cars, plumbing, house electrics, lawn mowers, washing machines, hi-fi; nothing was safe from my Dad and his two young sons. Inevitably it was off to do Engineering at Queen's University, from where I graduated in 1963. But well before 1963 I had found the conventional branches of engineering to be less interesting than I had wished. I headed off into Engineering Physics, great training for applied research postgrad studies. But in what? I had spent a couple of summers at the National Research Council in Ottawa, working in the radio astronomy group. It seemed to me at the time that astronomy was perhaps of passing interest and might offer decent engineering challenges. The astronomy got me in the end, but the engineering background paid rich dividends at various times in my later professional life. The immediate challenge was radio astronomy instrumentation, which I set out to do in a Master's degree program in the Department of Electrical Engineering at the University of Toronto, starting autumn 1963.

My joint supervisors were Donald MacRae, Professor and Head of the Department of Astronomy, and the brilliant and enigmatic J. L. (Allan) Yen, Professor of Electrical Engineering, theorist, instrumentalist, expert on Toronto Chinese cuisine (chopsticks were an early part of my graduate education) and a man who required almost no sleep. I saw both my supervisors but rarely, and then only when I was in trouble with them, this more frequently than was comfortable. I learned through the standard apprenticeship system, the senior grad students mentoring the new student intake. I learned most from Ernie Seaquist, who was well into his PhD program in the Astronomy Department. He was patient and generous to me with time precious for his own extensive radio astronomy program, and by example he taught me far more than just radio astronomy.

My project was to measure absolute temperatures of the galactic background at 320 MHz, using the pyramidal horn already installed at the David Dunlap Observatory (DDO), Richmond Hill, 19 miles north of Toronto. The horn itself (Figure 4.32) was in relatively good shape, needing some cleaning to remove certain avian deposits of the sort that Penzias and Wilson (1965a) encountered in their researches. The challenge as I mapped it out was (a) to build a reasonably low-noise amplifier and Dicke-switching receiver and (b) to design and build a reference cold load for the switching system, one with absolute temperature known to specified accuracy. The measurements were then simple drift scans, with the horn turned to the north celestial pole at periodic intervals for a reference level. This level would be calibrated by replacing the horn input with the reference cold load input. There were impedance-matching subtleties involved, as long-serving radio astronomers will recognize.

First task - to build a new receiver at 320 MHz. Field effect transistors, FETs, had just become available, actually working at this high a frequency! Low noise as well! But they cost real money, all of $34 each. In a rare interview with Allan, I got the money and the transistor. Next day I blew it up. (In retrospect I begin to understand the supervisor problem.) I managed to extract funds for a second one, and, after walking around it for an afternoon, made a decision on how to handle it which helped me the rest of my life. It's just another transistor! Handle with ordinary care - otherwise I couldn't see how I would get anywhere. It worked. I applied the lesson later when dealing with original astronomical plates. Treat them as you treat glass, with respect, but without awe. More tense and more "careful" = greater risk and less research.

The second FET ran throughout the project. The new receiver was built with help of George Watson, a solitary soul working out at Richmond Hill:

Fig. 4.32. The pyramidal horn antenna, aperture 3.7 by 2.8 m, used at 320MHz for my galactic background temperature measurements.

a craftsman, a perfectionist, and a delight, whose stories, unrepeatable and certainly unprintable, enlivened many of my days and nights in the little frozen cabin at Richmond Hill, while adding a certain breadth to my graduate education. More supervisor trouble ensued when in the course of transporting a frequency generator to the cabin (they weighed about 150 kg in those days), I settled the old radio astronomy station wagon axle-deep into the Observatory grounds in soft spring mud.

The cold load was a real challenge. Nobody really knew how to proceed, and the one I fashioned was the best technical achievement of my MSc. It did work well, and I was confident of its noise temperature - but note that it was a liquid nitrogen cold load, at about 80 K. This was close to the mean galactic brightness temperatures; but of course a long way away from CMBR values.

I heard/read of the CMBR as my observations progressed. Reaction (a): nothing to do with me; I'm a galactic (semi-) astronomer, working at too low a frequency and too high a mean brightness. Reaction (b), with minimal cosmic consciousness and from a radio astronomy point of view: surprise, Ryle was right after all - but a singular beginning? Steady state was conceptually much easier to handle.

And following this two minutes of deep thought, back to reality - the horn antenna had half-power beamwidths of 19.0° x 22.5°. Absolute temperature mapping requires correction for the response in side- and back lobes, of course. Thus I built a scaled version of the horn, complete with supporting structure, smaller by a factor of 9 and operating at 2.88 GHz. I mounted this on the antenna range turntable on the roof of the Electrical Engineering building, with a distant horn-reflector plus S-band generator to provide the signal. The main-beam and first side-lobe patterns agreed remarkably well with the main-beam measurements of the main horn using drift scans of the Sun, a point source (only 30 arcmin in size!) to the fat beam of the horn. The side and back lobes enabled me to estimate the spillover radiation.

There were many delays, including my MSc course load and stormy winter weather. Measurements began in February 1965 and continued to June; I covered the hottest part of the sky but by June (Figure 4.33), interference

L.S.T. : Local sidereal time

Fig. 4.33. A chunk of drift scan, this one at declination 5 = 40° complete with periodic visits to the North Celestial Pole and calibration-signal injections.

L.S.T. : Local sidereal time

Fig. 4.33. A chunk of drift scan, this one at declination 5 = 40° complete with periodic visits to the North Celestial Pole and calibration-signal injections.

from the USAF Buffalo base essentially halted the observations. I could not finish the cold (galactic anticenter) parts, another sore point between me and supervisors. My MSc thesis, complete with the iterative calculations to remove side- and back-lobe responses, was completed in October 1965. In parallel Donald Chu ran a sister set of measurements at 707MHz, using a 2.5-m precision horn-reflector at the Algonquin Radio Observatory of the National Research Council of Canada. The techniques he used followed mine precisely, including construction of a scaled model of the horn-reflector. His measurements and mine were to be used to calibrate in absolute terms higher-resolution galactic plane surveys at DDO with a new 10-m paraboloid reflector (for which I did commissioning and feed design.) These together with polarization measurements which Ernie Seaquist was working on were to provide comprehensive data on the Milky Way emission. This grander scheme never happened.

In November I set off for Australia, where I had been offered a scholarship at the Australian National University to do a PhD in a collaborative radio-optical program between Mount Stromlo Observatory and the Australian National Radio Astronomy Observatory at Parkes. John Bolton was to be my supervisor. My seduction by astronomy was complete. Engineering cropped up later in my life in building CCD systems, commissioning telescopes etc.; but it was astronomy now where my commitment lay.

Donald Chu, finishing the same patch of sky I had done, likewise left for different things, a proper job in his case with the then largest computer company.

In the excitement of starting a new life in a country where snow drifts across the telescopes were no longer a problem, the brightness temperature measurements were temporarily laid aside.

The rest of the story has a certain inevitability about it. Donald Chu had made some tentative comparisons of his data with mine; he found unsatisfactory answers. We knew roughly what the emission spectrum of the galactic background was - this synchrotron emission continuum from long-blown supernovae had a brightness spectral index of about —0.5 to —0.7 (Yates and Wielebinski 1967). Comparison of the 320- and 707-MHz results at independent map points by Donald and myself yielded a spectral index of —0.3, far too flat. Trying to reach indices in the "recognized" range meant zero-point errors outside our estimates. In 1965 we had left it at this: we had both moved on.

In 1967 or 1968, as cosmological consciousness dawned, I realized what had happened. Subtracting 3 K from both of our sets of measurements yielded spectral indices in agreement with the "known" results (Figure 4.34).

Spectral And Wilson

Frequency (MHz)

Fig. 4.34. The surface brightness measurements, circa 1969, from Wall, Chu and Yen (1970). PTS: Pauliny-Toth and Shakeshaft (1962); PW: Penzias and Wilson (1965a); HS: Howell and Shakeshaft (1966); RW: Roll and Wilkinson (1966); WCY: Wall, Chu and Yen (1970). ©1970 CSIRO Publishing.

Frequency (MHz)

Fig. 4.34. The surface brightness measurements, circa 1969, from Wall, Chu and Yen (1970). PTS: Pauliny-Toth and Shakeshaft (1962); PW: Penzias and Wilson (1965a); HS: Howell and Shakeshaft (1966); RW: Roll and Wilkinson (1966); WCY: Wall, Chu and Yen (1970). ©1970 CSIRO Publishing.

I collected the data together, redigitized it, and finally wrote up the experiments (Wall, Chu and Yen 1970). There was no great urgency at this stage.

In retrospect a dedicated CMBR measurement would have been simple. We had only to cover the colder parts of the sky, put our two sets of measurements together with a prior on the galactic emission spectral index, and a measurement of the excess radiation was there. We were a bit late in the time frame - but if we had got on with it in the first years of our MSc degrees rather than spending them wading through forgotten courses on plasma physics, the result would have been waiting for us.

The most astonishing aspect to me in hindsight was just how easy it would have been to make the measurement successfully, using the horns we already had, and a financial outlay of almost nothing.

I blame VLBI (partially). If Allan Yen had not become preoccupied with this (Broten et al. 1967) I know his razor-sharp mind would have seen the possibility; he read everything and was on top of everything. I know that excess radiation was in his mind - although he never mentioned CMBR or excess radiation to me, his annoyance when I had been unable to finish measuring colder parts of the sky convinced me of this. This too came in retrospect.

The CMBR subsequently played little part in my career of observational cosmology. I stuck to AGNs and their spatial distribution, together with schemes of (unified) beaming models. Most of this was with radio-selected samples. There were perhaps just three points of contact:

(i) In carrying out the (1984 version) deepest survey at 5 GHz with the VLA, Ed Fomalont, Ken Kellermann and I put limits on CMBR fluctuations in the range of an arcminute and a bit less (Fomalont, Kellerman and Wall 1984). These were the best upper limits at the time; but they were far from real detections at these angular scales, as we now know. Perhaps our main contribution was to determine how to minimize cross-talk between the antennas, a help to subsequent experiments. Even so, the VLA for all its power was never the instrument for CMBR fluctuations.

(ii) The standard model has the CMBR dipole, 1 part in 1000, explained as the Earth moving at 370kms_1 relative to the rest frame, with apparent temperature brighter in the direction of motion. The predicted motion should be visible in the number counts of distant objects, their combined surface brightness enhanced in the direction of motion of the Earth. There are serious difficulties in looking for this dipole in discrete objects: how distant, how to select, how to perform widescale calibration; what to do about obscuration, how to get beyond the cluster-dominated epoch. A uniform all-sky survey of radio sources offers hope, however, as Ellis and Baldwin (1984) pointed out. After completion of the superb NRAO VLA Sky Survey (NVSS; Condon et al. 1998), that hope could be really entertained. It took much work to understand the systematics of the survey, and much work to remove the nearby objects from it - but in the end Chris Blake and I succeeded in observing the dipole (Blake and Wall 2002), agreeing in magnitude and direction with Earth motion as implied by the CMBR (Figure 4.35). This remains the only detection of the velocity dipole in discrete galaxies, objects formed long after the epoch at redshift z ~ 1100 corresponding to the last scattering surface from which we see the CMBR. The mean redshift of our radio galaxies is about unity. The universe is therefore showing large-scale homogeneity at this epoch, and further analyses coupled with new deep and wide sky surveys can refine this result. Although few doubt the interpretation of the dipole in the CMBR, the detection in real

Fig. 4.35. Left: measured amplitudes of the deviation from mean surface density for NVSS sources, as a function of right ascension. (Note that the direction of the CMBR dipole lies - accidentally - close to the Celestial Equator.) The predicted amplitude is shown as the solid line. Right: error circles (1a, 2a) representing the direction of the NVSS dipole for samples selected at different flux-density levels. The point denotes the direction of the CMBR dipole.

Fig. 4.35. Left: measured amplitudes of the deviation from mean surface density for NVSS sources, as a function of right ascension. (Note that the direction of the CMBR dipole lies - accidentally - close to the Celestial Equator.) The predicted amplitude is shown as the solid line. Right: error circles (1a, 2a) representing the direction of the NVSS dipole for samples selected at different flux-density levels. The point denotes the direction of the CMBR dipole.

objects represents one of the tests the CMBR needs to pass if it is truly a relic of the big bang (Ellis 2002). (iii) With superb results from WMAP (Bennett et al. 2003), and with the Planck mission on the horizon, we would like some reassurance that the fluctuations we see in the CMBR are not contaminated by extreme inverted-spectrum populations of radio-millimeter sources. To this end, with Rick Perley, Robert Laing, Joe Silk, and Angela Taylor, I recently proposed a 43-GHz VLA survey of some 2 square degrees of the northern sky to search for such a population. This is the highest frequency search for extragalactic radio sources - and it found very few (Wall et al. 2006). We conclude that at small angular scales and the high frequencies of the measurements of the power spectrum of the angular fluctuations of the CMBR, there is little to fear from discrete radio source contamination.

I offer some conclusions.

(i) The CMBR was there all the time in our 1965 data; and we could have done the measurements earlier with specific attention to detecting it as a part of our absolute flux measurements. It would have come in somewhere between 3K and 5K at a guess. I think it's a stretch to say that we would have believed it on its own; our frequencies were a little low. But had there had been contact with cosmologists such as between Penzias and Wilson (1965a) and Dicke et al. (1965), then it might have been different. Too if my cosmic consciousness had not dawned so slowly, it might have been different.

(ii) I cannot have any regrets. My MSc project was superb for starting research in observational astronomy. How better to learn everything about the basics of radio astronomy? Every aspect in the process was revealed to me in glaring detail, all the pitfalls, noise, bandwidth, line-loss, mismatch, spillover, ground radiation, antenna patterns, conversion of antenna temperature to brightness temperature It was baptism by fire, and I did love it, I think. It is next to impossible for a student nowadays to learn about instrumentation in depth at any wavelength, and I grudge a big vote of thanks to my supervisors Donald MacRae and Allan Yen for so comprehensively dropping me into it.

(iii) It is possible to observe and not see. After all, Donald Chu and I were only a couple of engineers playing around with horn antennas . . .

Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment