The Objective:
The product as tested failed to meet the EMC standards. The objective of the investigation by ETS was to identify the cause of the excessive emissions and implement a fix. It was very important to the customer that the cost of introducing the fix would not effect the production cost of the unit. This constraint is common in massed produced items and dictates a solution based on suppressing noise at source rather than more expensive shielding and filtering further from source.

The Investigation:
The pump controller consists of two PCBs in a screened steel enclosure. Ribbon cables connect the PCBs to each other and power and control cables link the controller to external devices. The first stage of the investigation was to identify which PCB was the source. The board loop antenna of the ETSi 4334 was placed over each PCB in turn and a simple level measurement made of the emissions at the problem frequency. This quickly identified the main CPU board as the culprit.
The next stage of the investigation was to pin point the source of emissions at this frequency to a track and component level. To do this, the PCB was placed on the ETSi4334 and a 1mm X-Y plot carried out at the offending frequency. This took less than 20 minuets. This finished plot contained sufficient detail to enable the PCB designer to immediately identify two problem areas; A track from the CPU to a peripheral IC and a hot-spot beneath a driver chip. If necessary, the gerber file for the component layout or track layers could have been overlaid on top of the X-Y plot to assist with identification.

Plot of emissions from original board

Armed with this level of information on the problem it was a simple matter to implement suitable suppression. Components costing only a few pennies/cents were applied to suppress the driver IC and filter the signal track.
The suppressed PCB was re-scanned and compared with the original scan to confirm the effectiveness of the fix. This showed a significant reduction in the noise level as shown. The system could then be re-assembled and tests on the OATS completed.

Plot of emissions from modified board

Conclusion:
The total time to identify and suppress the fault including dismantling and re-assembly was less than two hours. The system went on to successfully meet the applicable standards, and the cost of the extra components was acceptable for production.