The module will be offered without a package and should be a genuine all-round device. A number of potential circuit environments thus have to be taken into account:
- Is the printed circuit board housed in a metal or plastic package?
- Is the circuit ground well connected to the package in the metal version?
- Is the circuit operated in the vicinity of other metal conductors such as a top-hat rail, 230 V mains lead, etc.?
The module has to be protected against all possible interference mechanisms since its future environment is unknown.
Disturbances may enter the module via conductors and/or fields.
We will initially consider conducted interference: disturbances may enter the module via the power supply socket (a switched-mode power supply unit, for example) or the peripherals’ interface (a temperature probe, for example). The magnetic fields caused by disturbance currents flowing through the board may induce voltages in the conductor loops. Two problems have to be taken into account with regard to safeguarding the module’s functionality: the induced voltage may either be treated as a logic signal by the integrated circuit’s input or it drives a disturbance current, which causes problems in other parts of the integrated circuit.
All conductor runs have been relocated in the printed circuit board’s intermediate layers to prevent this. Critical signal pins of the controller such as the reset pin and the sockets which connect the printed circuit board to the outside are fitted with filter elements.
The same correlations apply to disturbances which enter the module via fields. Magnetic field vortexes may penetrate the circuit and induce a voltage in the conductor loops, which in turn drives a disturbance current through the module and causes the aforementioned problems.
Interferences are also caused by electric coupling. Electric fields capacitively couple into the circuit board’s line networks or even components. The resulting displacement current may cause a voltage drop at a resistor (against Vss or Vdd), which in turn is recognized as a logic signal, or induce voltages in other parts of the circuit.
The bottom of the printed circuit board that is only populated on one side is provided with a continuous GND layer as a counter-measure. The top is also GND-flooded to minimize the disturbing influence of magnetic fields.
Both capacitive and magnetic coupling also have to be considered on the integrated circuit level.
H-field coupling causes a magnetic field vortex to penetrate the IC. A disturbance voltage is induced in the IC’s current loops. The induced voltage may interfere with signals or the supply voltage in the IC and cause faults or drive a disturbance current through the conductor loop and thus interfere with the integrated circuit.
During E-field coupling, a voltage which is present between the IC and field source generates an electric field depending on the respective IC-to-field source distance. The electric field lines end on the IC’s metal parts (pad of the IC pin, bond wire, die). They conduct a displacement current into this surface (Figure 1).
Since the integrated circuit’s EMC itself cannot be influenced, a controller has to be found with the highest possible immunity for the application. A number of integrated circuits with a comparative range of functional features are potential candidates for this application. The manufacturers’ data sheets, however, do not reveal the respective immunity parameters. A new criterion has thus to be found to evaluate the immunity with E and H field coupling.
The aim is to evaluate/compare the potential integrated circuits in terms of their immunity to disturbances coupled in via fields.
Either one or both coupling mechanisms (H-field/E-field) can cause faults depending on the IC’s design. An objective immunity evaluation thus has to subject the integrated circuits to disturbances via both coupling mechanisms.
The chosen approach is shown in the Figures 3 and 4.
Separate coupling circuits were designed for both coupling mechanisms. An EFT/burst generator (burst generator according to IEC 61000-4-4) was used as a disturbance source. This generator was connected to the coupling waveguide via a 50 Ω high voltage cable for H-field coupling. The wave guide had a 50 Ω input to ensure that the burst reaches the device under test without distortion. An additional measuring shunt monitored the generated disturbance pulses.
The wave guide was arranged above the devices under test at the defined angle and distance. This guaranteed that all ICs were subjected to a comparable disturbance field with an identical EFT/burst generator setting.
The pulse shape generated by the waveguide (measured via the shunt) is shown in Figure 3.
Apart from the generator setting, the magnetic field’s angle also had to be taken into account since it is related to the waveguide’s orientation and has a direct influence on the interference effect achieved.
A similar set-up was chosen for E-field coupling (Figure 5).
The EFT/burst generator was connected to a coupling electrode instead of a waveguide. This electrode was arranged at a defined distance above the device under test. The voltage between the coupling electrode and the device under test generated an electric field proportional to the burst voltage amplitude.
The devices under test could thus be subjected separately to E-field and/or H-field disturbances.
The subsequent measurement was expected to show that the individual integrated circuits fail completely or cause faults at different EFT/burst generator voltages, coupling mechanisms (E-field/H-field) and field angles. Or in other words: the integrated circuits’ immunity to E/H-field disturbances differed from manufacturer to manufacturer and the measured results let the engineer choose a suitable IC for the described application.
80C51 microcontrollers from three manufacturers were examined as potential candidates for the application in the course of the measurements described below.
The integrated circuits were not tested in the application but on customized test adapters to create reproducible conditions and prevent parasitic effects by other parts of the circuit.
The following parameters applied for the measurement:
- Identical package pin-out (VQFP44)
- Comparable functionality – all three are 8051-compatible
- Identical test adapter (packaged with the same filter elements)
- Identical test program and/or firmware
Preparation
The integrated circuit was tested during operation. The test program was selected so that each component in the integrated circuit (timer/UART/watchdog, etc.) was used and the corresponding test signals on the pins provided information about their functionality.
A pin was continually toggled (heartbeat signal) and a static signal sent to the outside in the present example. An oscilloscope was sufficient to monitor this test set-up. In addition, the outputs were connected to LEDs to receive visual feedback about the operating state of the device under test. The individual operating states of the IC were controlled by a PC via a test adapter-to-PC connection.
The test program ran in the following way: LED_01 (heartbeat) flashed slowly while LED_02 came on permanently during the start of the IC. Depending on its firmware, the IC changed over to another operating mode which caused LED_01 (heartbeat) to flash faster and switched LED_02 off should a crash and subsequent reset occur. Irregularities of the heartbeat signal indicated an internal program sequence problem.
The subsequent figures show the measurement set-up used for the test procedure (Figures 7 and 8).
It comprised the following components:
- EFT/burst generator with a maximum generator voltage of 4.4 kV
- Base plate for the test adapter with an integrated IC-to-PC interface
- Device under test in the test adapter
- H-field source/E-field source with a 3 mm spacer
- Oscilloscope and oscilloscope adapter
- Power supply for the PC interface and IC
The IC was connected to the PC interface via the test adapter. This allowed the engineer to monitor and control the IC.
The measurement set-up shown in the figure also included an oscilloscope adapter which made it easier to connect the oscilloscope’s scanning heads and did not affect the measurements.
A controlled switched-mode power supply unit with an internal current limiting function supplied the measurement set-up with power and was intended to protect the IC from destruction in the event of a malfunction.
The field sources were connected to the EFT/burst generator. The 50 Ohm measurement output of the field source was connected to the oscilloscope to monitor the injected pulses.
The measurement procedure was as follows: first of all, the IC’s program was started. Its proper functioning was monitored by the oscilloscope. The field source was placed over the centre of the integrated circuit, starting with the H-field source.
It was important to adjust the field source relative to the device under test when this was subjected to an H-field. This ensured that the results are comparable since the interference effect depends on the field angle. The field angle did not have to be adjusted if an E-field was coupled in.
Field coupling was started at the EFT/burst generator’s lowest amplitude value and a positive polarity. The severity was then gradually increased up to a maximum generator voltage of 4.4 kV or until a fault occurred. The polarity was then switched over and the measurement repeated. Several measurements had to be taken at different field angles under the influence of H-field. The integrated circuits were subjected to the disturbance for one minute in each of the test runs.
Tables 1 and 2 summarize the results and show at which voltage amplitude, polarity and field angle a reset occurred for the different IC’s.
| Manufacturer | Polarity | Generator voltage at the moment the circuit failed | ||
| Angle 0° | Angle 90° | Angle 180° | ||
| Manufacturer 1 | positive | 2,040 V | No failure | No failure |
| negative | No failure | No failure | 2000 V | |
| Manufacturer 2 | positive | 4,000 V | No failure | No failure |
| negative | No failure | No failure | 4000 V | |
| Manufacturer 3 | positive | 3,600 V | No failure | 800 V |
| negative | 800 V | No failure | 3,300 V | |
Table 1: Immunity level determined during H-field coupling
| Manufacturer | Polarity | Generator voltage at the moment the circuit failed |
| Manufacturer 1 | positive | No failure |
| negative | ||
| Manufacturer 2 | positive | |
| negative | ||
| Manufacturer 3 | positive | |
| negative |
Table 2: Immunity level determined during E-field coupling
Only three field angles were chosen for an initial test of the ICs under the influence of a magnetic field. A second test run with a finer resolution can be carried out to pinpoint any functional faults that occur.
The integrated circuits’ different immunity levels become visible straight away. None of the tested integrated circuits was susceptible to E-field.
A crash could only be invoked in the 80C51 IC from Manufacturer 2 at 4 kV while the IC from Manufacturer 3 carried out a reset at a value as low as 1 kV when subjected to magnetic field. Since all of the test conditions (test set-up, interconnection, test program, etc.) were identical, the differences must be inherent to the integrated circuits themselves.
The measurements at a field angle of 180° provided the same results as the measurements at a field angle of 0° with the opposite generator polarity.
The deviations which occurred and are clearly visible in the table can be explained by variations within the generator. These can be verified on the basis of the pulse shape generated at the oscilloscope’s measurement output.
None of the integrated circuits could be influenced at a field angle of 90°.
The heartbeat signal was not influenced during any of the measurements, i.e. the IC’s were functional until the reset. In view of these findings it seems reasonable to assume that the integrated circuits’ power supply was disturbed.
Figure 9 shows the top view of a device under test with a spacer. The spacer has a degree scale where the position of the waveguide (H-field) and thus the field angle can be read.
The Vcc and Vss pins are on opposite sides of the IC package in the present example. A maximum voltage is induced in this loop at a field angle of 0° and/or 180°, leading to an IC power supply failure and thus a reset.
Since no other faults occurred and none of the integrated circuits was susceptible to E-field, the generator voltage at which the ICs failed when subjected to H-field was used as a comparison criterion. As a result, the 80C51 from Manufacturer 2 was chosen for the application since it has the highest immunity level of the ICs measured. After this EMC assessment the engineers can proceed to the development of the modular electronic switchgear.
Dipl.-Ing. Lars Glaesser studied ele
