Note: This is an excerpt from the book, EMI Troubleshooting Cookbook for Product Designers, by Patrick André and Kenneth Wyatt. See Reference 1.
Introduction
Radiated emissions are by far your highest risk when performing compliance testing at the test facility. With all the high-speed digital circuitry inside electronic products today, it becomes all too easy for harmonics of the clock frequencies and other fast-edged devices to be the source of radiating EM fields. Typically, failure modes will be cable radiation or leakage from enclosure seams or apertures.
Here is a handy checklist you can use, either as a pre-test check prior to compliance testing, or as a check following a test failure. Because the “clock is ticking” at the test lab, the checklists and recommendations are brief and to the point, so you avoid wasting too much time.
Radiated Emissions Check List
- Frequencies below 200 MHz, likely the cables are the source of radiation. The wires are better antennas at lower frequencies, where the wavelengths are longer.
- Frequencies above 200 MHz may come from the chassis. The higher the frequency, the more likely it is the chassis of the unit, or a circuit board when there is no chassis or an open frame unit.
- Assure that all shielded cables have a low impedance bond at both ends. Assure the shields are terminated with direct contact to the chassis or connector. Avoid using a pigtail unless absolutely necessary.
- If using a pigtail to bond a shielded cable, assure it is as short as possible.
- Assure chassis metal pieces are making excellent contact with each other (10 milliohms or less) – no paint or other coatings, grease or dirt, corrosion or oxidation, which could create an impedance.
- Verify that each line leaving the equipment is filtered, and that the filter is located next to the point of penetration out of the equipment.
- If performing commercial testing (FCC, CE, etc.) and vertically polarized emissions exist below 80 MHz, try lifting the power cord away from contacting the ground plane. This will reduce the coupling path from product to antenna via the ground plane. Conversely, try increasing contact with the ground plane and see if emissions increase.
- If there is support equipment connected to the unit, assure it is not the source of the noise. Turn off the support equipment if you can. If not, turn off your equipment and leave the support equipment on. If the signals remain, your source may not be the equipment under test, but the support equipment.
Typical Failure Modes
Most products fail the radiated emissions test due to radiating cables or leaky chassis enclosures. Here are some things to check quickly.
Cable Radiation — I/O or power cables generally radiate high-frequency harmonics due to poor bonding of the shield to chassis or enclosure, lack of adequate filtering or are simply “poked through” the shielded enclosure. Generally, if you have failures below 200 MHz, that indicates cable radiation. The reason lower frequency emissions tend to come from the cables is the need for physical length to make a good antenna (the bigger, the more efficiently they can transmit emissions). The cables tend to be the longest part of the equipment, and thus the source of most low frequency emissions.
Metal Chassis — Higher frequency (typically greater than 200 MHz) emissions are more common from the metal chassis of the equipment. At these higher frequencies, the I/O cables tend to become inductive and therefore are higher impedance than the chassis is for flowing RF currents, and so, tend to radiate. One exception to this is when the equipment being tested is physically large. A seven-foot high metal cabinet, which is resting on the ground plane, may have a quarter wave resonance around 30-40 MHz.
Seams — One common source is the seams of the chassis enclosure. Circuit boards inside the unit can generate currents on the inside surface of the chassis. These HF currents leak out of seams or gaps, and then will flow around the outside of the chassis or enclosure of the equipment. Thus, the entire enclosure becomes a transmitting antenna. An exception is when most of the current can be returned to the source very close to the point where it is coupled onto the chassis. This is why it is good to use bypass capacitance on a circuit board, or circuit board reference return planes which are well bonded with the chassis.
However, when HF currents flow inside the enclosure of the equipment, and they come to a seam, they must be able to flow across that joint very easily. Impedances of a few milliohms will create a voltage across the seam (a strong E-field) that can radiate. Note that a horizontal seam will have a voltage gradient or vector from top to bottom, creating a vertically polarized E-field and a vertical seam will produce a dominant horizontal polarized E-field. A good troubleshooting technique is to note the polarity of the E-field (assuming an E-field antenna is used) and determine if this could be generated by a poorly bonded seam.
If the product includes an LCD display, leakage can occur around the edges of the display. In many cases, LCD displays are not well bonded to the enclosure and the entire display can act as a transmitting antenna. Other areas of leakage can include the spaces between plug-in daughter cards (as is used in the typical PC chassis) or ventilation ports.
Troubleshooting Emissions at the Test Lab
- Often you will need to troubleshoot the emissions at the test laboratory. There are several things you must be aware of and be able to do.
- You must be able to see the display of the spectrum analyzer. This may be a projected image inside the chamber, or an analyzer, which is brought into the chamber. If the only option is a monitor which is placed at the door of the shield room, and the door must be left open to see it, be sure the emission you are observing is not an ambient signal from FM broadcast, cell phones or digital TV. You may need to turn your equipment off to assure this is true.
- When you observe the emissions, realize you are loading the room just by being in the chamber. The emissions will not be the same level as before. Also, the azimuth of the unit under test may not be at a maximized position (for commercial testing). However, this may change by slight movements of the cables and unit, so be aware of this fact when you feel an improvement has been made – it might have only shifted in angle or position.
- Do not stand between the antenna and the equipment. The human body makes a wonderful radio frequency absorber.
When approaching a radiated emissions problem, and once the setup is as shown, make sure you are in control of the area. Tell people where to stand and make sure they don’t move. Their movement and position will affect your investigation.
Start by grabbing cables with your hands (if it is safe to do so). If the cables are the main radiators, you may be able to identify a cable quickly by grabbing and releasing the cable. While you do this, minimize your movement, since your movement will also tune and detune the room and area. You may need to take a wood or plastic stick to lift cables without coming in contact with them and to minimize your influence on the cables. A hockey stick may be used for this purpose, which allows you to stand some distance from the unit, minimizing your effect on the radiated field. Thus, any changes you see will be from the cable movement.
Many times, I/O cables may be connected to the product under test, but are disconnected at the far end. Try disconnecting these cables one at a time, leaving them off until all cables not used during the testing have been disconnected. This may also help identify which cable, or cables, are radiating.
If the cables do not seem to be the problem, try placing your hands on the case or chassis of the equipment, again only if it is safe to do so. Press and squeeze the box if possible to assure metal pieces are touching, or to open their contacts. In this case, you may see the emissions make a sudden jump up or down, indicating a make or break of contact somewhere. While you are there, look for coatings or overspray that might be between metal interfaces.
Assure the support equipment is not at fault. Turn it off, if you can, without turning off significant portions of your equipment under test. If that can’t be done, attempt to reverse that, and turn off the equipment under test, while leaving the support equipment on. Does the problem remain? If so, the support equipment may be the issue. This is also true if the support equipment is outside the chamber. The cables passing from outside to inside can contain a significant amount of RF energy, which can rebroadcast inside the chamber. Assure these cables are well filtered, shielded, or somehow treated to avoid this issue. Sometimes, loading these long support cables with a series of several ferrite chokes can effectively “remove” them from affecting the actual system or product under test. If you cannot power down either the equipment under test or the support equipment, try changing loads, states of operation, data rates, or other functions and watch for changes on the emissions.
It may be helpful to have a non-conductive plastic or wooden crochet hook. Use it to pull individual wires out of a cable bundle. If safe to do so, you can touch these wires with your fingers to determine if they are sensitive, changing the emission levels as you pinch and let go of the wire.
One of the best ways to identify radiating cables is to measure the common-mode currents traveling on the wires or cable shield. By clamping a current probe with spectrum analyzer around them close to the unit you can measure the RF current in the wires, which is strongly correlated to radiated emissions. In fact, for electrically short cables (less than a quarter wavelength), it’s possible to predict the e-field in V/m, which may be compared with the regulatory limits. This will be described shortly.
Consider purchasing a pair of long aluminum knitting needles. Wrap one of them most of the length with insulation tape (e.g. black electrical tape). You can use this to probe the connectors, connector pins, circuit boards, cases and chassis parts, by touching them with the conductive end of the knitting needle (be cautious about shorting connector pins, etc.). Watch for increases or decreases in emissions while you do so. Both of these can identify a sensitive area, which should be investigated more carefully. Instead of knitting needles, you can use a disconnected multimeter probe, or a connector pin with a wire soldered onto it. In fact, these may be easier to use, since the wire can be oriented in the same direction as the antenna polarity while connected to the sensitive area.
Emergency remedies for small to medium sized equipment include wrapping the whole unit in aluminum foil. Since large areas need to be covered, it is best to not use copper or aluminum tape. Also, aluminum foil does not suffer from the build up of impedances the way conductive tape does. That is to say, when using metallic tapes with conductive adhesives, remember that the adhesive is not very conductive, as much as we would like it to be. As you layer tape on tape, the resulting build-up of impedances can greatly reduce the effectiveness of the shield. Using aluminum foil without any coatings will improve this bond by orders of magnitude.
To use, aluminum foil in this manner, fold the seams of the foil over several times, as in the seam of your pants. Bond the foil to any connectors and cable shield possible. To assure bonds, use wire ties or “zip ties” around the connectors. If this continues to radiate, try placing the foiled unit on the conductive ground plane (yes, the floor, if that is serving as the ground plane). If it continues to radiate, then likely the cables are still an issue.
If this cures the problem, then the chassis is likely at fault. Slowly peel back the aluminum foil over areas you feel are less likely to be an issue, such as over solid panels without displays or connectors. Reveal connectors and displays last. Each time you peel back some foil, check to see if the emissions are returning or if they remain low. Often this is best performed while watching a monitor of the emissions while you perform the work.
Summary
Testing your product for EMC compliance at third-party test labs can be a harrowing experience. Radiated emissions is usually the riskiest test and I hope these suggestions might be of benefit in your future testing. For additional suggestions for the remaining EMC compliance tests, I’d refer you to Reference 1. Finally, for best success, I’d also refer you to the article in this guide, How to Prepare Your Product and Yourself for EMC Testing.
References
- André and Wyatt, EMI Troubleshooting Cookbook for Product Designers, SciTech Publishing, 2014.
- 2017 EMC Precompliance Test Guide, Interference Technology.
