Derby and District Astronomical Society
The Journal of the Derby and District Astronomical Society
May - August 2006
Radio Astronomy Experiments at 12 GHz
By Tony Razzell
Normal amateur astronomy uses a tiny fraction of the available electromagnetic spectrum, between about 0.4 and 0.7 microns wavelength. The first of the 'New Astronomies' that appeared in the 20th century was radio astronomy, at wavelengths less than ~20 metres. The first amateur radio astronomer was Grote Reber, a ham radio enthusiast who built a 30ft dish in has back garden in Illinois, and produced the first ever maps of the sky at radio wavelengths. Since Reber’s pioneering work, the development of electronics has been enormous in the intervening 60 odd years. Satellite television has resulted in the availability of very sensitive microwave receivers at very low cost, which are comparable with some that the professionals were using in the 60’s and 70’s. A number of amateur radio astronomers have built radio telescopes using surplus satellite TV equipment, so I thought I’d have a go at doing the same.
A simple amateur radio telescope (known as a total power instrument) consists of an antenna to collect radio waves from space, a low noise amplifier to increase the signal strength whilst adding minimal noise, a frequency converter to drop the signal down to a lower frequency, an IF (intermediate frequency amplifier), and a detector to convert the signal to a DC voltage. An IF amplifier is used as it is easier to build high gain amplifiers at lower frequencies (this receiver arrangement is known as the superheterodyne or superhet). The low noise amplifier and frequency converter are usually built into the same unit for satellite TV, and sit at the focus of a small dish antenna. A schematic arrangement of the telescope is shown in Figure 1.
Figure 1. Simple total power instrument
The first problem was the microwave receiver and dish antenna. I acquired my first receiver (known in the satellite TV trade as a Low Noise Block or LNB) for £5 at an amateur radio rally. This was followed by a rather bent Sky TV dish and LNB found by a friend of mine in a skip. I then came across a really good 0.9 m dish with an offset feed, mounting and LNB which was dumped next to the lane along which I go to and from work. A Sky minidish (newer technology than the old Sky TV unit) also turned up dumped in the lane, so that was collected as well. Having got a selection of LNBs, I needed something to plug them into. An LNB converts the microwave signal from the satellite frequency ~12Ghz (or 2.5 cm wavelength) into a lower frequency (0.9-2.1GHz) that is used by the satellite tuner that sits on top of the TV. What I needed was an equivalent unit to the tuner to act as the IF amplifier, but which converted the signal into a voltage that I could record or read with a meter. The book on Amateur Radio Astronomy by William Lonc is a very good source of information, as well as several sites on the internet. Eventually, I decided to make my own IF amplifier, based on application notes for MAR-6 MMA devices and the book by Lonc. The circuit diagram is shown in Figure 2.
Figure 2. IF amplifier circuit diagram
The IF strip uses monolithic microwave amplifiers (MMAs) which provide wide bandwidth amplification at 1-2GHz with minimal effort (not having a wide range signal generator, the actual bandwidth is unknown but is assumed to be at least 500MHz). An inductor is used as a radio frequency choke to supply power to the LNB without losing the signal coming back (the same cable is used for power supply and signal). The signal from the LNB is then fed into the first of three MMAs which are linked together to produce a gain of ~60dB. Appropriate decoupling and short leads must be used at these frequencies, and I made a circuit board with a copper ground plane on the back for stability. The third MMA is followed by an envelope detector. I used a circuit from a web site by Michael Fletcher (http://www.qsl.net/oh2aue/fig20.gif), which uses a Schottky diode, plus identical reference diodes to compensate for temperature variation (a simpler single diode circuit could equally be used). Variation of gain and therefore output with temperature is one of the main practical issues faced by the would-be radio astronomer (see later). A potentiometer is used to adjust the DC offset so that the cold sky can be set somewhere near zero. An alternative (used by many beginning amateur radio astronomers) is to use a dish alignment meter or a signal strength output from the satellite TV tuner if it has one.
When the IF strip was complete, I connected an LNB, a power supply and a meter on the output. The sun was very easily detectable on the old Sky TV dish/LNB, as well as strong signals from the various satellites that live in an arc about 38° above the horizon in the south. My main dish is the offset fed 0.9 m as this captures the greatest incoming signal and has the narrowest beam width. The Sky minidish produced a huge signal, even compared to the 0.9 m dish, which appeared to have a high quality German LNB on it. To record signals properly, I used an ancient, but serviceable Acorn Archimedes 310 computer with an analogue input board (essentially mimicking the BBC Micro model B I/O facilities). A separate DC amplifier was used to increase the output voltage from the IF strip unit from a few tens of millivolts to a few volts. A data logger programme was used to record data on the Archimedes, which was easily transferable to a PC as a comma separated variable text file on a floppy disc.
To record an object in the sky, I used a transit method where the dish was pointed at a fixed altitude and azimuth slightly to the west of the object to be observed, and the output from the radio telescope recorded as the object drifts through the beam. In this way, the earth's rotation scans the sky and the signal from the object shows as a peak in the recorded signal strength. The 0.9 m dish has a beam width at the 3 dB points (50% power) of ~2° at 12 GHz. At an altitude ~38° the rate of drift is ~1° every four minutes, so a scan of ~30 minutes is enough to capture the sun as it drifts through the beam of the dish. A transit curve from the sun is shown in Figure 3 (it was found that the peak amplitude varied somewhat from trace to trace, probably due to pointing inaccuracies).
Figure 3. Transit observation of the Sun
A problem immediately noticed was drift of the output voltage. This is caused primarily by changes in temperature, which varies the gain of the instrument. This is illustrated in Figure 4, where the telescope was left running for several hours, with an accompanying increase in output voltage, exceeding the peak from the Sun drifting through the beam. As can be seen on the x (time) axis, this drift was during the late afternoon/early evening accompanied by a drop in temperature.
Figure 4. Drift due to change in receiver gain caused by temperature drop
A more difficult object to detect is the Moon. This was achieved, but the resultant signal is similar in magnitude to the thermal drift of the instrument, and fairly noisy (Figure 5). The spikes at the beginning and end of the trace are the ~310K blackbody radiation coming from a hand placed over the mouth of the LNB in an attempt at providing a crude calibration.
Figure 5. Raw Moon transit observation
An attempt was made to correct for thermal drift by fitting a notional straight line between the background start and end points and subtracting this from the data. The corrected plot is shown in Figure 6. The reason for the flat top during the transit is unknown, the 'shoulder' is assumed to be short term drift.
Figure 6. 'Corrected' transit of the Moon
A combined plot of the moon and the sun is shown in Figure 7 to (roughly) the same amplitude scale, offset to give approximately the same background sky level.
Figure 7. Combined Sun and Moon transits to approximately the same scale
Future improvements to the instrument will concentrate on improving gain stability. One way of doing this is to periodically switch the input from the antenna to a stable reference signal. In this way, any drift in gain can be subtracted from the wanted signal, eliminating the baseline drift effect seen in Figure 4. At 10GHz, this is not a trivial task, and I may resort to a mechanical approach such as periodically putting a lossy absorbant material in front of the LNB horn. This arrangement is known as a Dicke switch, and was an early development in radio astronomy. A more elegant solution would be to electrically switch the LNB input from the horn to a reference (or dummy load) using diodes or transistors. This would be preferable as the reference could be kept at a stable temperature. I have modified an old Sky TV LNB connecting one of the two inputs to a resistor to act as a dummy load (the units have two inputs to allow for polarised signals from the satellite), and experiments with this are ongoing. Another approach would be to put the electronics in a temperature stabilised box, which I am also considering.
I have enjoyed building the 12 GHz telescope, and found it fascinating to see blackbody radiation from the Sun, Moon and even my own body. The satellite equipment at 12 GHz is very cheap (free for most of mine) and easy to set up to detect the Sun and I intend to publish updates in future issues of the British Astronomical Association (BAA) Radio Astronomy Group's journal Baseline. [and hopefully Aries - Webmaster].