RSP RFI
Posted: Tue May 24, 2016 9:15 pm
Numerous articles and YouTube videos have been published on how to deal with RFI issues. Most proposals end up proposing creative shielding designs or by asking for a metallic casing. With limited reliable results.
RF interference manifests itself as a higher noise floor and as spurious signals. To analyze these issues, the RSP has been taken to an EMC test range.
The measured results highly depend on the test frequency and the setup of the test configuration. All tests were performed at a frequency of 149MHz. The length of the cables connected to the antenna port was 80cm. Different results may be obtained for other frequencies and setups, e.g. in SW bands.
How do unwanted signals enter the receiver?
The most common way how spurious signals may appear is by loading the receiver with strong signals from the antenna. Less efficient antennas, filtering or even a resistive attenuator may help. But this is a topic for a different discussion.
The cables connected to the antenna and the USB port form a dipole antenna. The signal picked up between the ground terminals of these cables spreads over the board and finds its way to the receiver. This received signal has been measured to be >40dB lower compared to a real antenna placed at the same position.
A good and proven practice to keep this antenna signal from entering the receiver is to add a choke (ferrite) to the antenna, to the USB cable or to both. The choke can be placed quite elegantly within the casing over the cable connecting the antenna port and the board.
The chokes do not keep the USB signals from entering or leaving the board, however. These signals are bound within the shielding of the cable.
A really efficient solution (to be considered for future designs) is to short the ground terminals of the antenna and the USB connector with as low an impedance as possible. Both the connectors should be close together and not at opposite sides of the casing. Connecting the ground of the two ports results in a signal attenuation of >15dB.
The board itself may also receive signals. The parasitic antennas on the board are quite inefficient. The received power is >50dB below that of a real antenna. Hence, this effect is of less importance.
Due to the limited shielding effect of coax cables, the antenna cable itself may pick up RF signals. These signals will not be suppressed by the chokes placed over the cable. The signal received via this path has been measured to be >30dB lower compared to a real antenna. These tests have been run with a well-terminated RG58U cable of length 80cm. This effect may appear for strong close-by signals entering the antenna cable. The shielding effect of coax cables has not been measured yet at lower frequencies, such as those of noise emitted by switching regulators. The effect could be worse there.
In summary, the most important impact will be from signals picked up by the dipole antenna formed by the cables connected to the antenna and the USB input port of the receiver. These signals have to be blocked from entering the receiver. Signals penetrating the coax shield may also result in RFI. The effect of signals entering directly into the board is of minor influence. For the given test, setup no effect on the noise floor has been observed.
RF interference manifests itself as a higher noise floor and as spurious signals. To analyze these issues, the RSP has been taken to an EMC test range.
The measured results highly depend on the test frequency and the setup of the test configuration. All tests were performed at a frequency of 149MHz. The length of the cables connected to the antenna port was 80cm. Different results may be obtained for other frequencies and setups, e.g. in SW bands.
How do unwanted signals enter the receiver?
The most common way how spurious signals may appear is by loading the receiver with strong signals from the antenna. Less efficient antennas, filtering or even a resistive attenuator may help. But this is a topic for a different discussion.
The cables connected to the antenna and the USB port form a dipole antenna. The signal picked up between the ground terminals of these cables spreads over the board and finds its way to the receiver. This received signal has been measured to be >40dB lower compared to a real antenna placed at the same position.
A good and proven practice to keep this antenna signal from entering the receiver is to add a choke (ferrite) to the antenna, to the USB cable or to both. The choke can be placed quite elegantly within the casing over the cable connecting the antenna port and the board.
The chokes do not keep the USB signals from entering or leaving the board, however. These signals are bound within the shielding of the cable.
A really efficient solution (to be considered for future designs) is to short the ground terminals of the antenna and the USB connector with as low an impedance as possible. Both the connectors should be close together and not at opposite sides of the casing. Connecting the ground of the two ports results in a signal attenuation of >15dB.
The board itself may also receive signals. The parasitic antennas on the board are quite inefficient. The received power is >50dB below that of a real antenna. Hence, this effect is of less importance.
Due to the limited shielding effect of coax cables, the antenna cable itself may pick up RF signals. These signals will not be suppressed by the chokes placed over the cable. The signal received via this path has been measured to be >30dB lower compared to a real antenna. These tests have been run with a well-terminated RG58U cable of length 80cm. This effect may appear for strong close-by signals entering the antenna cable. The shielding effect of coax cables has not been measured yet at lower frequencies, such as those of noise emitted by switching regulators. The effect could be worse there.
In summary, the most important impact will be from signals picked up by the dipole antenna formed by the cables connected to the antenna and the USB input port of the receiver. These signals have to be blocked from entering the receiver. Signals penetrating the coax shield may also result in RFI. The effect of signals entering directly into the board is of minor influence. For the given test, setup no effect on the noise floor has been observed.