Home - Techniek - Electronica - Radiotechniek - Radio amateur bladen - Ham radio - Understanding computer-generated RFI
Some remedies for this malady
Radio Frequency Interference (RFI) has always been a source of concern for the Radio Amateur. Until the recent introduction of complex circuitry in the home (like microcomputer systems, VCRs, and microwave ovens), our communications gear has been the main source of potential RFI. In this computerized age of ours, increasingly complex and sensitive receivers are being barraged by these and other sources of RFI.
Microcomputers have become an integral part of the contemporary Amateur Radio station. Virtually all modern receivers and transceivers rely on microcomputer-controlled circuitry for their internal operation. But more important, from the perspective of RFI generation, is the external standalone microcomputer system. This system can be found in an ever-increasing number of Amateur Radio stations running programs for predicting HF propagation, logging and checking QSOs, printing QSL cards, and even controlling transceivers. Unfortunately, many computer systems radiate a significant amount of RF into the shack. This is especially true on the HF bands, where even a small amount of RFI can mask an otherwise readable signal. This article examines the microcomputer as a source of RFI, and suggests some steps you can take to contain it.
To understand how the microcomputer can be a source of RF interference, you must have some knowledge of how microcomputers are constructed, how they operate, and how they are normally connected to other devices.
On the most basic or physical level, most microcomputers are composed of a system unit, power supply, keyboard, and display screen. Depending on the make and model of the microcomputer, these components may be physically separate, as in the IBM PC, or found in a single package, as in many portable computers. Many microcomputers include the power supply and at least one floppy disk drive in the same physical enclosure as the system unit. Other models have the power supply and disk drives packaged separately. Many of the popular microcomputers have slots for extending the basic system with plug-in modems, memory cards, and video cards for a variety of screens.
Few of us make do with the minimal system configuration. The most common additions are printers, extra disk drives, and modems. Adventuresome hams have packet controllers and interfaces to their microcomputer-based transceivers. All of these additional devices, and the cables that connect them to the system unit, are potential sources of RFI.
For the purposes of this discussion, consider the microcomputer to be composed of a Central Processing Unit (CPU), memory, and a system clock. The CPU is the heart of the microcomputer. It not only performs operations on data that resides in memory, but also keeps track of the current status of the executing program and handles communications with memory and input/output devices like printers and modems. All of this activity must be carefully orchestrated for the microcomputer to function properly. The metronome for this activity, and a potential source of RFI, is the all-important system clock.
The system clock creates the timing signals used to synchronize all activities within the computer. And as in our modern transceivers, a quartz crystal normally serves as the basis for this timing. In some microcomputer systems the CPU contains the oscillator circuitry, so an external crystal is simply connected between two pins of the CPU chip. In other systems, separate dedicated chips are used to generate the timing signals. In the IBM PC/AT, for example, there's a clock generator chip and a programmable timer chip. The clock generator chip, which uses a quartz crystal, creates the basic timing signals used by the computer.
The programmable timer chip is related to the clock generator chip. Think of it as a programmable array of flip-flops that produces an output signal every so many clock cycles. For instance, if the basic clock cycle is 6 MHz (as in the original IBM-AT), and you want to perform some event every 1/60,000 second, the programmable timer chip can be programmed with a count of 100. At every 100th clock cycle, the programmable timer chip will produce a signal that can be used by the computer circuitry.
| Apple IIE/C | 1 or 4 MHz |
| Apple Ilgs | 1 or 6 MHz |
| Commodore 64 | 1 MHz |
| Compaq Deskpro-286 | 8 MHz |
| IBM AT | 6 (original) and 8 MHz (later) |
| IBM PC | 4.77 MHz |
| IBM PS/2 Model 30 | 8 MHz |
| IBM PS/2 Model 50 | 10 MHz |
| Macintosh 512/Plus/SE | 7.8 MHz |
| Macintosh SE-30 | 16 MHz |
| Macintosh II, IIx, IIcx | 16 MHz |
Obviously, RF energy at both the clock frequency and the variable programmable frequency (6 MHz and 60 kHz in the example above) are of concern to the Amateur. See Table 1 for a listing of clock times used in the popular microcomputer systems. In some cases, the basic crystal oscillator circuitry operates at the same frequency as the system clock. For example, the 10-MHz IBM PS/2 Model 50 uses a 10-MHz crystal. In other instances, a crystal oscillator with a frequency higher than the clock frequency is used in conjunction with frequency divider circuitry. For example, the IBM PC uses a 14.32-MHz crystal with a clock frequency of 4.77 MHz (14.32 divided by 3).
Most microcomputers with built-in power supplies, like the Apple Macintosh and IBM PC/AT and PS/2 series, make use of internal lightweight switching power supplies. Switching supplies, unlike conventional linear supplies, do not make use of large iron core power transformers. Instead, the AC from the 110-volt power line is directed to a bridge rectifier, and the resulting ripple DC is pulsed at between 20 and 100 kHz. This pulsed, high frequency DC allows for the use of small, lightweight high frequency transformers. The pulsed DC and its harmonics are potential sources of RFI, through the power lines and cables to computer accessories. Also, whereas the relatively massive, high inductance power transformers effectively block RF radiation into the power lines, the high frequency transformers used in switching supplies can potentially couple RF more easily into the AC power line.
Sooner or later most of us will add an electronic keyer, power amplifier, or beam antenna to our bare-bones rig. Similarly, those of us bitten by the computer bug are seldom satisfied with a minimally configured microcomputer. The most common additions include floppy and hard disk drives, various types of printers, modems, and alternative input devices. Depending on your computer's design, many of these additions may take the form of cards that plug into the system, or external peripherals that must be connected by cables to the system unit. In order to understand how these devices can cause RFI, you must have a basic grasp of their operation.
Although the keyboard is by far the most common method of interacting with the computer, there are various alternative input devices. What follows is a short description of the most common ones.
The Mouse: Some micros, like the Apple Macintosh, come factory equipped with this cursor control device. Other systems, like the IBM PS/2 series, have a mouse interface which lets you select a mouse of your choice. From an RFI perspective, the mouse, which is the most popular alternative or supplement to keyboard entry,(1) can be classified as mechanical or optical. The mechanical mouse uses a roller ball that moves as you control the mouse. Two perpendicular rollers (one for the x-axis and one for the y-axis) attached to contact pins are coupled to the roller ball. These contact pins make and break connections with a contact bar as the mouse moves, much like a distributor in an automobile engine. Optical mice are similar to mechanical mice in many respects; however, there are no physical make-and-break connections. They use LEDs and phototransistors to detect motion. As you might expect, the rapid make-and-break connections associated with mechanical mice can result in RFI. In comparison, optical mice are more electronically "clean." But, like any peripheral attached to the system via a cable, the mouse cord can act like a broadcast antenna for signals inside the system unit.
Trackballs: Trackballs are best thought of as inverted mechanical or optical mice. They offer the same benefits and limitations as the other mice in terms of their potential for RFI.
Light Pens: Light pens work by sensing the exact time that the electron beam in the monitor excites a phosphor at a particular point on the screen. The associated circuitry of the light pen determines the x-y coordinates of the point on the screen by measuring the time it takes for the electron beam to reach the pen. Light pens pass this information back to the computer through a cable, or in some cases, by sending an RF signal to a receiver mounted on the top of your monitor.(2) In directly wired systems, light pens generate low intensity signals only every 1/60 of a second. Assuming that the cable is adequately shielded, the potential for RFI is relatively low. RF light pens, in comparison, have a high RFI potential.
Tablets: Graphic tablets, useful for drawing and tracing, come in three basic types: electromagnetic, resistive, and acoustic.(3) The most RFI-prone type, the electromagnetic version, has a handheld pen that transmits an RF signal to a receiving grid located under the tablet surface. Tablet circuitry converts the signals into x-y coordinates to determine the pen's exact location.
Resistive tablets, sometimes called touch pad tablets, are made of two conductive surfaces separated by a small air gap. When a pen touches the tablet, bringing the two surfaces together, current flows between the two surfaces. The strength of the current is used to determine the x-y coordinates of the pen.
Acoustic tablets use a pen transmitting ultrasonic waves (65 to 75 kHz) that are received by microphones near the work area. Through triangulation, the relative strength of the received signal at each microphone is used to calculate the relative x-y coordinates of the pen.
Touch screens: Touch screens, like light pens, are useful for selecting objects on the screen. The most common varieties are mechanical, optical, and capacitive. Optical screens use rows and columns of infrared LEDs - photo-transistor pairs mounted opposite each other along the edges of the screen. Touching a particular point on the screen with your finger blocks one or more x-y beams of light. The touch screen sends the coordinates of the broken beams to the computer, which calculates the corresponding x-y location on the screen.
Mechanical switching panels are composed of transparent, conductive membrane switches mounted over the display screen. Pressing your finger on the panel brings the two conductive surfaces together and completes the circuit. One sheet determines the x-axis and the other the y-axis location of the contact.
Capacitive touch screens have a capacitive coating on the CRT screen that acts as one plate of a capacitor. When you make contact with the screen, current flows into your body from the contact point. Sensors on the screen detect the location of the current drain and calculate the corresponding x-y location. In my experience, capacitive systems are more prone to RFI than either the mechanical or optical versions.
Joysticks: Although they are more popular as a game interface than a way of manipulating Amateur software programs, joysticks should not be overlooked as a source of RFI. The vast majority of joysticks are mechanical, corn-posed of switches and/or potentiometers. The largest RFI threat from these simple devices occurs when the connecting cable acts as a radiator for system unit signals.
Modems (named for MODulator-DEModulator) let digital computers communicate over analog phone lines. In some microcomputer systems, like the Apple II series and the IBM PC/XT/AT, modem cards can be inserted easily into slots in the system unit. Along with minimizing the potential for RFI, these internal modems have the added benefit of providing a less cluttered ham shack. The more RFI-prone external modems, also popular on the IBM PC and other microcomputers, must be connected by a cable to the system unit. Purchasing an internal modem doesn't guarantee freedom from RFI, however, because the telephone cable represents a potential RF antenna.
Printers range from simple dot-matrix units to complex laser printers that contain their own RF-producing microcomputer systems. Although a few printers attach directly to the system unit (primarily on portable models), the vast majority are connected to the system via cables. In my experience, the mechanical printers are less likely to cause RFI, but the acoustic noise they produce is hardly bearable during a QSO. The relatively silent laser printers, by comparison, emit considerable RF energy.
Many of the lower priced microcomputers are designed to work with TV receivers as monitors. RF modulators con vert the video signal into a VHF signal that can be handled by a TV receiver (commonly on channel 3 or 4). These so-called "RF bricks" are potential sources of RFI - especially when the output is unshielded 300-ohm flat line.
A local area network lets microcomputers communicate with other devices connected to the network, including modems, printers, computerized communications gear, and other microcomputers. ICOM's system allows multiple ICOM receivers and transceivers to communicate with each other and with microcomputers. Because this system directly connects your computer system to your communications gear, there's ample opportunity for RFI. In my experience with the ICOM system, there's no detectable RFI as long as the integrity of the cable and associated connectors are maintained. Other networks, like Apple's AppleTalk, can cause considerable RFI. AppleTalk is a low speed network often run over standard (unshielded) telephone cable. Receive on my ICOM 751A is rendered practically useless when AppleTalk is in operation.
There are a number of steps you can take to minimize computer-generated RFI in your shack. In some cases, your problem may be cleared up by following only one or two of these measures. In more difficult situations, you may have to try all of these suggestions for acceptable results.
Check the FCC rating before you buy: Microcomputers are rated by the FCC as either Class A or Class B devices, depending on the amount of RFI produced by the equipment (see Table 2). Paradoxically, the often cheaper Class B machines, intended for home use, are less prone to RFI than the Class A machines. The more expensive Class A micros, including many micros based on the Intel 80386 chip and several of the large monitors, have less stringent RFI ratings. Unfortunately, many hams use the more powerful and more RFI-prone "business class" machines at home.
| Maximum radiation measured at 3 meters | ||
|---|---|---|
| Class A | Class B | |
| 30 - 88 MHz | 3.000 µV/m | 100 µV/m |
| 88 - 216 MHz | 5.000 µV/m | 150 µV/m |
| 216 - 1000 MHz | 7,000 µV/m | 200 µV/m |
| Maximum conduction into the AC power line | ||
| 0.45 - 1.6 MHz | 1000 µV | 250 µV |
| 1.6 - 30 MHz | 3000 µV | 250 µV |
Think twice about a micro with a Class A FCC rating. Taming a machine with a Class B rating is difficult enough! Check the Model/Serial number tag on all computer equipment, including modems, printers, scanners, and mice before you buy. You should see an FCC ID number, together with a statement like "Certified to comply with the limits for Class B computing device pursuant to Subpart J of Part 15 FCC rules" These rules are designed to provide "reasonable" protection against RFI in a residential installation.
Shielding: Don't defeat the shielding in your computer or peripherals. You may be tempted to remove the aluminized plastic backing from the Macintosh SE or Commodore 64 motherboard to prevent heat buildup. Don't! You will have a very cool-running RFI machine. Also, keep the rear metal card covers on your system unit intact. If you remove an internal card, make certain that you replace the original slot cover.
The new IBM PS/2 machines, like the original Macintosh series of computers, make heavy use of metalized plastic for shielding. Unlike the original PC and many of the PC clones these new lightweight machines limit metal shielding mainly to the power supply. If you aren't careful when opening these plastic cases, you might chip or wear away the conductive paint, and have a less than perfect RFI shield.
Cables: Use shielded cables whenever possible, and add snap-on ferrite inductors to peripheral cables - especially if they aren't shielded. You can use ferrite snap-on toroids to increase the series inductance of cables, raising their impedance to HF signals. Although adding significant inductance may have the effect of reducing the computer signals, the high frequency RF components will be attenuated to a greater degree. Peripherals (like external disk drives) that can create their own RF signals should have ferrite snap-on inductors attached to both ends of the cable. The inductor near the peripheral attenuates signals generated from within the peripheral, while the inductor near the computer system attenuates signals generated by the system clock that may be inadvertently coupled to the peripheral "antenna." Don't forget to add an inductor to the telephone cable where it exits your modern.
Bypassing: Judicious use of RF bypass capacitors with resistive touch pads, mechanical mice, and joysticks often pays off.
Power Conditioning: The simplest way to provide a good degree of isolation between your computer equipment and your receiver is to make certain that each system is controlled by different circuit breakers. Plug your computer and peripherals into a wall socket that is not connected to the socket used for your communications gear.
If using separate power circuits fails to remedy your RFI problem, or if all of the sockets in your shack are controlled by a single circuit breaker, try adding two good surge protectors to your shack - one for your communications gear and one for the computer equipment. A simple protector with MOVs won't do. The best method of isolating signals coupled through the power line uses a combination of RF line filters and transient suppressors. You can realize a 60-dB attenuation of interference above a few hundred kHz with RF line filters.(4)
Ground: Although you have no doubt heard it before, a good ground is essential for minimizing RFI. It's surprising how many hams who have 6-foot ground rods connected by heavy coaxial braid to their gear fail to ground their computer equipment. Treat your computer system, including all peripherals, like your communications gear where ground is concerned, and you should be well on your way to minimizing potential RFI.
Layout: Minimize cable lengths. When possible, use an internal modem instead of an external one with its associated cables. If RFI persists, try rearranging your equipment. Move your micro and peripherals as far from your receiver as possible. In some cases, interference can be minimized to acceptable levels through proper layout of equipment and judicious cable runs. Obviously, running your external disk drive cable parallel and adjacent to the antenna feedline is asking for trouble.
Communications Gear Modifications: Try to minimize the number of possible entry points for computer-generated RFI into your system. If you have an external speaker with more than a few inches of cable, use a low pass filter and shielding to prevent the speaker wire from acting as an antenna. Software Design: If you develop your own software, try to minimize the reading and writing of data to disk. Similarly, when you purchase software developed by others, run the program and make note of how often the disk drive whirs. The stepper motors and associated drive circuitry are extremely noisy in the RF spectrum.
The best way to handle computer-generated RFI is to think of your computer system as you would any other piece of RF communications gear, with peripheral cables, phone connections, and power cords acting as the antenna system. Use low pass filters on all antennas (snap-on toroids on all cables and power cords), make certain that you provide a good system ground, use shielded cables of minimum length, and start with "RF-clean" gear. Use bypass capacitors whenever possible, and keep the computer "antenna system" away from your communications gear.
NU1N, Bryan P Bergeron.