SM 5 BSZ - Static rain.
(July 13 2011)


Sometimes raindrops get charged when they fall through the atmosphere. They may become charged to many kilovolts with respect to ground. When such a raindrop approaches a grounded conductor such as an antenna element there will be a discharge which induces a significant pulse voltage into the antenna feedpoint.

A raindrop might have a diameter of 4 mm. That means it has a capacitance to infinity of 0.2 cm. That is about 0.2 pF. If the potential is 10kV an archover will happen when the droplet is something like 3 mm away. The plasma that results is a good conductor and therefore essentially an inductor. As a result the current will continue after the potential on the raindrop has reached zero and the droplet will charge to the opposite sign. After a while it is likely to have moved closer and then a new arcover will happen. Each raindrop will cause a series of discharges according to this theory. I can not say for sure that the theory is correct, the plasma could also be essentially a non-linear resistor.

Real world data.

The practical result is for sure a sequence of pulses for each raindrop. See figure 1 which is a 144 MHz recording sampled at 2 MHz.

Figure 1.Static rain recorded with a Perseus at a sampling rate of 2 MHz.

The first pulse is stronger than the subsequent ones. The pulse repetition frequency decreases with time. The number of pulses and the ratio between the first pulse and the remaining pulses varies from pulse to pulse, but not by more than aboout a factor of two.

Pulses of this kind can be totally eliminated with the smart blanker on a properly calibrated system. This recording is however made on a system that is far from properly calibrated.

In this particular situation there are no strong signals present so the interference can sucessfully be removed with the dumb blanker.

Figure 2 shows the same situation with a 20 times lower sampling speed. Here two channels are sampled at 96 kHz with the WSE converters connected to a X-yagi.

Figure 2.Static rain recorded with WSE converters at a sampling rate of 96 kHz.

Figure 2 with a sampling rate about 20 times lower than figure 1 does not resolve the pulses in their individual components. As a consequence the pulses do not have a shape that is determined by the frequency response of the system and the smart blanker is totally useless despite the fact that the system is properly calibrated.

The dumb blanker can be used, but with heavier static rain, the pulses would appear too often.

The pulses appear far more often in channel 2. This is evident in figure 3 which shows the S-meter vs time for the two channels.

Figure 3.Static rain recorded with WSE converters at a sampling rate of 96 kHz. The pulses come in groups. Far more frequently in channel 2 which is the horizontally polarized yagi. Some droplets hit the vertical elements, but far less often. The bandwidth is 5 kHz only so individual droplets are not resolved

The fact that the pulses come in groups is well known from the sounding of the QRN when it is starting or ending. Whether this is true during heavier rainfall is unknown to me at this time.

Missing information.

To extract useful information about static rain a bandwidth of at least 1 MHz is required. There are two aspects of interest to me.

  • 1. How much does the pattern (number of sub-pulses and separation in time between sub-pulses vary from time to time? What are the highest pulse rates one should expect?
  • 2. How well can Linrad remove the static rain noise under difficult circumstances with several very strong stations in the passband?
  • 3. Is there a similar phenomenon static snow?
Recordings that shed light on those questions are most welcome. I will provide an ftp site where large files can be uploaded.


The screen dumps on this page are from files supplied by Carsten Gabriel, DM1CG.

To SM 5 BSZ Main Page