Local Problems for RadioJove at ISU
Joachim Köppen DF3GJ Kiel/Strasbourg/Illkirch Spring 2005
The worst and most persistent problem in 2003/04 was that of a periodic noise interference which commenced when the street lamps were switched on in the evening, and stops when they were switched off in the morning. Thus, any night time observations for Jovian emission are seriously hampered. A typical emission looks like this in the raw data:
To the left, there is the sequence of calibration steps, but to the right the almost strictly periodic increase of noise is apparent. During the interval of high signal one also notes an unpleasant sharp hissing in the audio signal which is unlike the hiss from a noise generator. The period is measured as about 47 seconds, there is some minor variation of this value. Since we noticed that one of the sodium vapour street lamps in front of the ISU building was often seen flashing, a study revealed that it did so with the same period. Some nights, the level of interference showed variations in strength, sometimes even it stopped for some minutes. This interference was not present in the first days in january, but hope that the problem had solved itself were shattered with the resumption of the emission. Later, the periodic behaviour changed into a almost chaotic pattern, since one or a couple of noise bursts were left out. When we were sure that it was indeed thia lamp, we called the local council to repair the lamp, and we have not had any trouble ever since.
With the calibration, we can give the noise levels in temperature (this plot has also been processed to remove narrow spikes and noise fluctuations, as explained below)
The emission thus switches from a low state of about 100 kK to a high state of about 200 kK. We note that the expected background from the MIlky Way is about 50 kK, with Jovian signals about 10 to 30 times higher.
In spring 2005 we saw a similar pattern, much waeker and lasting only a few days. Apparently the repair crew came early!
The spikes on the plots caused by short radio impulses, created by electric discharges and computers, can be removed by software, in the SkyPipePlotter program. The technique is to check every datum whether it lies above a straight line connecting data 10 pixels apart ("width"); if it does so by more than a factor (1+height) of the current y-value, the datum is replaced by the value predicted by the straight line. Smoothing is done simply by averaging over the "width" pixels left and right of each pixel. One fiddles with these three parameters, until one obtains what one accepts as a nice result... To show that such a software can fully remove narrow spikes, I display the observations of a solar burst, taken by John Rock in England:
The raw data, shown in grey, is first processed to remove the narrow spike, then we also smooth it to remove the fluctuations on the signal from the noise. The resulting curve is shown in red.
Recently, I came across another phenomenon: As I was listening, a noisy
mechanical, throbbing sound - like the awful noise used by jamming transmission
during the cold war period - came up and faded again in a fairly regular
way, with a period of a few seconds. Since transmissions with this sound
are often heard on short waves at various frequencies - but not in the
broadcast bands! - they are obviously a kind of data transmission, though
I do not yet know the proper name.
This transmission I could hear independent of the tuning of the Jove receiver, so it obviously was a strong signal on a nearby frequency which caused overload of the mixer. The ups and downs in its volume were simply the almost periodic fading which occurs usually when the reflectance of the ionosphere changes as its ionisation decays at dusk or builds up at dawn. The signal was probably only a few 100 kHz away from the centre of the filter's passband, so whenever the signal became strong, it could leak through into the receiver...
The plot showed a regular series of rather broad peaks which one could have identified as interesting solar bursts:
Starting in winter 2004/05, the sudden rise of the noise level was noticed on some days, sometimes in the morning. The noise kept its high level (about 2 million K) until it suddenly came off again, sometimes in the evening. Clearly, the origin is a machine that is switched on and off at certain times. What initially appeared to be a pattern associated with normal working hours, did not persist. The noise was high during some nights, and once it lasted a couple of days. However, normally there are many days and nights without this noise. Unfortunately it was always turned on whenever there there good opportunities to listen to Jovian bursts...
On one day, the noise can be seen to be switched on to a first level at 12:41, then it became even higher at 14:28; later it was switched off again in two steps, at 18:17 and 18:21. The narrow spikes are distant lightning.
Below is shown a signal which sometimes is observed, at any time of the day, and with various speeds: It is a radio carrier sweeping through the frequency range more or less quickly. If it is more than 20 kHz below or above the receiver frequency (20.17 MHz), it is not mixed into the receiver passband, thus inaudible. But as soon as is comes nearer, it is mixed into the audio frequencies and it shows of in a rise of the tracing. First the audible tone is of high pitch, but comes down in frequency, until the tone has dropped below some 100 Hz, which is below the lower limit of the audio passband. Then it becomes again inaudible - hence the tracing drops also. When the signal's frequency is exactly the same as the one from the receiver's oscillator, the mixing results to zero frequency ("zero beat") which is the centre of the dip in the M-shaped tracing. After passing thtrough this point, the audio frequency rises again and becomes audible, and finally it becomes again inaudible, since the signal now is too far away from the receiver's oscillator frequency to produce a audible tone. When one listens to the receiver, one may pick up one of these events. What the purpose of these signals might be, we don't know. They do no come in a regular fashion, so they can't be swept-frequency measurements of the ionosphere...
| Top of the Page | back to Main Page | back to my Home Page |