Each new design can introduce unfamiliar failure modes, but take solace, there are common failure types that can be reasonably avoided or subsequently managed.
Let’s Knock Out the Common Issues First
Some of the most common sources of EMC failure in new designs are worth understanding before your product reaches the test house.
Noisy power conversion: many suppliers offer tools to simulate their devices. Also look for reference design EMC results and adhere to recommended layouts.
Crystals: ensure they are presented with the required impedances, are as close to their load (processor for example) as possible, and even consider guard traces or rings in your layout if there is a history of emission concerns.
Clocks and clock harmonics radiating: clocks are such reliable emitters they are a handy way for test houses to ensure a product is operational and coupling to the measurement equipment. Don’t forget the impedance matching of long traces and cables applies to the harmonics too! Bury clocks between ground layers, if possible, or take caution managing the return path proximity on PCB ground layers or cables.
Shielding: think complete shielding or nothing. Shielding is expensive and can act as a reflector if not incorporating a clean ground plane AND the traces from inside to outside are filtered or treated for inadvertent radiation/reception.
Cabling: the number one offender is still a necessary part of supply, communication, or sensing. However, it is often the most overlooked in a new design from an EMC perspective. Cheap cables can work, providing the twist rate is consistent in twisted pairs or the connectors don’t separate the forward and return paths so much as to cause mode conversion or create loop antennae.
Characteristic impedance performance is a major factor in higher frequency communication, often a big cost differentiator, especially for longer distance applications. Remember that degraded functional performance equates to reduced immunity and higher susceptibility.
Insufficient protection circuitry: remember, the protection circuitry needs to be activated by the transient energy, which requires a certain amount of time for the transient waveform to rise, unimpeded, before it activates the nonlinear circuitry. Then, consider the redirected energy flow, is it still close enough to disturb my sensitive circuitry?
Failures by Test Type
- Transient Immunity Failures
The protection device fails to redirect all the high current flow during a discharge. This can manifest itself in many ways:
(a) The protection device might not be switching on quickly enough, exposing the sensitive circuitry to part of the transient disturbance.
(b) The abrupt transition from the protection device switching can cause a secondary switching spike leading to damage.
(c) The redirected energy flow is still too close to the sensitive circuitry.
- Conducted Immunity
Weaknesses are often observed in lower frequency ranges and can be quite challenging as the design may need to balance competing filter cutoff frequency requirements. At higher frequencies, narrow failing bands are generally easier to treat. The causes can be more obvious where the interfering signal might line up with lower speed communication protocols or clock frequencies. The less obvious issues are mode conversion.
The test setup might be injecting a common interference signal on a pair of lines, and the product manages to direct that energy in a balanced form for most of the test frequency range. However, any imbalances present can cause a differential signal to directly compete with the communication or even offset DC levels. The good news is that it’s much easier to identify the band and the cause when troubleshooting in the presence of a live failure, compared to high-voltage transient failures.
- Radiated Immunity
Probably the most difficult because the precise lab conditions of this test are not easily reproduced anywhere else. Radiated immunity failures are most often caused by cables receiving and conducting the interference into the device under test. Reproducing the failure in a conducted fashion is a great way to allow for easier troubleshooting.
- Radiated Emissions
These are often the clearest failures to diagnose but require careful consideration to resolve. Identify the emission source, which is often obvious if you know the product well, then focus on locating the antenna, the traces, cabling structures, etc. that allow the source to radiate efficiently. Paying close attention to antenna orientation (horizontal or vertical) helps pinpoint the product’s inadvertent antenna orientation and, from there, the true source of the emissions.
- Conducted Emissions
My favorites are low-frequency failures, which are often tangible and repeatable, making them easier to probe and trace. Focus on identifying the source quickly and determine early on whether it’s common-mode or differential-mode emissions.
Let’s Talk About Damaged Components
If a circuit smokes, explodes, reddens, or is no longer attached to the PCB after testing, appreciate the information it has offered. While this may be unsettling, it can lead to faster troubleshooting and accelerate the redesign effort. A component, track, or connection has likely been subjected to more current or voltage than it was designed to handle, making it the obvious starting point for investigation.
After you have reassured yourself that the energy should have gone through this part, consider that it’s rating coincides with what you expected the transient energy to have been. After replacing the part, testing at lower levels to just warm the part (not destroy this time) can be helpful too, especially if you have a thermal imaging tool (ask your lab about it).
Figure 1: Thermal Image Scanner for Identifying Circuitry Warming Up
Know What to Avoid
While this is all quite general, future blogs will provide more depth. The hope here is that these examples are what you have always known (and managed) to avoid.
Read all the blogs in the EMCguy series