Go to Top

EMI in output/input connectors


Connectors suffer from many EMC problems due to they can be considered as short length conductors with a rigid body. Sometimes it is difficult that manufacturers ensure a 360º shielding in the connectors because it depends on the many factors such as case shielding, number of grounded pins or spring length.
Thereby, connectors are an important factor to take into account from the EMC design point of view even if they are not in use because they can amplify the internal system noise to outside like an antenna. In order to avoid this issue, connectors should be properly shielded and a possible solution could be to place sealed caps in the open interfaces. Thus, the effectiveness of shielded caps is going to be evaluated in this document through testing a device which holds a D-SUB connector due to rectangular connectors often show more EMC problems.
Many devices have D-SUB connectors to interface computers or systems with others peripherals. A D-SUB connector have two or three parallel rows of contacts generally surrounded by a metal shield with “D” shaped that provides mechanical support and protection against electromagnetic interference. Nevertheless, this metal case sometimes does not avoid interference on this type of connectors when they have nothing connected. Therefore, the objective of this test is to evaluate the D-SUB-RF-SEALED Cap solution manufactured by Würth Elektronik and designed to protect against RF influence in this kind of interface.


A spectrum analyzer capable of measuring signals from 10 Hz up to 3 GHz and a Würth Elektronik near field probe has been used to evaluate the D-SUB-RF-SEALED CAPS in a system designed to generate radiated and conducted emissions in a metal enclosure. The test will be held by attaching the near field probe close to the connector and measuring the noise emitted in the full spectrum analyzer frequency range. Next, the test will be done with the cap placed in the connector.
Both measurements will be acquired with the max hold option enabled in order to compare them and to able to find the noisiest frequency range. Once this range is detected, it will be analyzed with more detail by zooming in that frequency range.
An electromagnetic scanner module has been also employed in this test to evaluate the effectiveness of D-SUB caps. This is formed by a precise near field probe arrays and it is able to detect the magnetic field information in all directions with the spectral / spatial scanning feature. It can scan a frequency range for the selected region and measure the radiation situation in the whole frequency band of each spatial location to get the full electromagnetic information.
The device that will be under test is the EMC Box which generates high-frequency pulses through an internal oscillator circuit. The output signal is connected to 2 pin of the D-SUB female connector placed in the box front side. The connector case is joined with the shielded metal enclosure and the oscillator circuit GND plane. This device allows characterize female connector taking it as the worst case due to pins of male connector usually generates more emissions.


The EMC measurement environment of the design laboratory is very difficult that meets the ideal conditions to do this test due to there are uncontrolled reflective surfaces, such as metal furniture, walls or cable trays. Furthermore, there may be electronic equipment and electromagnetic noise which could couple in the measurements. Therefore, a reference measurement has been done in order to measure the laboratory electromagnetic noise.


Figure 1. Reference measurement: electromagnetic lab environment.

The setup employed to carry out this second experiment is based on the D-SUB EMC Experiments Box connector. The test has been performed with the internal oscillator on and with the external cable unplugged.


Figure 2. Measurement method in the EMC Box

The next spectrum analyzer capture shows significant noise peaks different from the ambient noise up to 2 GHz and a large attenuation in some of these peaks when the cap shielded is placed.


Figure 3. Measurement of the D-SUB connector without cap (yellow) and with cap (blue) in the 10 Hz to 3 GHz frequency range.

It is possible to analyze these peaks better in the next capture which shows the 10 Hz to 1 GHz frequencies. The mains peaks are attenuated from 3 dB up to 20 dB and others have been suppressed such as 200 MHz and 800 MHz peaks.


Figure 4. Measurement of the D-SUB connector without cap (yellow) and with cap (blue) in the 10 Hz to 1 GHz frequency range.

With regard to the 1 GHz to 2 GHz frequency range, the main peaks also have been attenuated when the D-SUB cap is placed in the connector. For instance, 1200 MHz peak has been suppressed or 1600 MHz peak that has been attenuated more than 10 dB.


Figure 5. Measurement of the D-SUB connector without cap (yellow) and with cap (blue) in the 1 GHz to 2 GHz frequency range.

Next, the effectiveness of D-SUB-RF-SEALED Caps will be evaluated with the EMSCAN to locate better the noise emissions origin in the connector. In order to carry out this measurement process the next diagram has been implemented.


Figure 6. EMSCAN connection configuration

The setup employed in this test is based on placing the front side of the EMC Box on a determined area of the EMSCAN as shown in Figure 7.


Figure 7. Measurement method

The test has been performed with the D-SUB connector in the row 20 and the columns 5 to 8 so that it is not necessary to evaluate all the scanner area. It should be noted that in that position the number scheme of the D-SUB connector is inverted, so the pin 2 is placed in the cell 20-7. Thereby, only the cells close to the D-SUB connector will be analyze in order to reduce the scanning time. Figure 8 shows the selected area.


Figure 8. Scanned area

With regard to the frequency range and acquiring parameters, the test has been performed by selecting a start frequency of 10 MHz and a stop frequency of 100 MHz due to the main noise components are concentrated in that range. In order to obtain a great balanced between scan time and resolution a RBW of 30 KHz has been set.
The first test carried out shows the spectral of the signal measured in the selected EMSCAN area. By viewing the spectral capture, the maximum peak corresponds to 404,8 MHz, with an amplitude of -88,3 dBm that is to say, 26 dBm higher than the base noise level. Another data that provides this capture is the special position of this measure in the EMSCAN: row 20 and column 7.


Figure 9. Spectral of D-SUB electromagnetic noise measurement without shielding.

These data can also be analyzed by viewing the spatial radiation diagram as shown in Figure 10. It is possible to identify the cells with more electromagnetic noise, and where is placed the pin 2 of the D-SUB connector where is connected the oscillator circuit output. It is identified by the red point and is in row 20 and column 7 as indicate the spectral graph.


Figure 10. Spatial radiation diagram without shielding.

EMxpert PC application allows the user to adjust the reference noise level, so that, it this level is increased up to -105 dBm the higher radiation is focused on the right side of the D-SUB connector as illustrates Figure 11. If the reference level is moved to -95 dBm, the noise source is highlighted even more, that is to say, the pin 2 is easier to identify in the diagram. Furthermore, the application indicates which is the amplitude level and in which frequency has that point.


Figure 11. Spatial radiation diagram without shielding increasing the reference noise level.

It is also possible to show the acquired data in a spatial 3D model to evaluate the area which is influenced by the radiation. In this case the amplitude is represented by the vertical axis.


Figure 12. Spatial 3D radiation model without shielding.

Once the electromagnetic noise of the EMC Experiment Box D-SUB connector has been studied, the D-SUB-RF-SEALED CAP is placed in the connector and the same measurements will be repeated to evaluate its performance.
Now, the spectral capture shows a large reduction of the maximum peak. If the last maximum peak located in 404,8 MHz is compared after placing the shielded cap, it is possible to view that the amplitude level has been attenuated 17,2 dBm. Figure 13 illustrates the spectral measured with the shielded cap placed in the D-SUB connector and the maximum frequency peak.


Figure 13. Spectral of D-SUB electromagnetic noise measurement with shielding

If the spatial radiation diagram is studied, the same attenuation effect can be observed due to the maximum noise point is in row 20 and column 7. This point corresponds with the pin 2 of the D-SUB connector where is connected the oscillator circuit output but in this case the radiation is more distributed around the studied cells.


Figure 14. Spatial radiation diagram with shielding

In order to observe better the radiation level in the spatial diagram, next figure shows both captures with and without placing the shielded cap using the same noise reference level, -105 dBm. The green area had been strongly reduced when the shielded cap is plugged in the D-SUB connector and the red point is much smaller than the data measured without shielding.


Figure 15. Spatial radiation diagram with (right) and without (left) shielding increasing the reference noise level.


After measuring some D-SUB connectors in several systems the main noise peaks are summarized in order to obtain conclusions about the attenuation rate:


Therefore, the effect of D-SUB-RF-SEALED Caps has been evaluated in this test through measuring the electromagnetic noise emitted in a D-SUB connector. The EMC Experiments box based on an oscillator circuit placed inside a metal enclosure has been used as noise source and it has been connected to an external D-SUB connector.
Considering the measurements taken with the EMC scanner and the Spectrum Analyzer before and after placing the shielded cap, both measurements shows an attenuation higher than 10 dBm in the maximum peak placed around 400 MHz in the EMC Box tests. Furthermore, it also has demonstrated a large attenuation in the rest noise peaks in the frequency range analyzed (10 – 2000 MHz).