Screened Room Outperforms OATS for EMC Tests on HV Power Supplies
Graham Muncey, Start Spellman, UK; John White, Rainford Systems, UK -- Test & Measurement World, 8/1/1999
While many companies initially consider that only complete systems need to meet EMC requirements, there is now increasing acceptance that system components should also comply. Today, suppliers of both systems and system components recognise the scope of EMC requirements, and this recognition has increased EMC testing in small- and medium-sized companies. Start Spellman (Pulborough, UK) is one such medium-sized company that recognises that their products need to meet current EMC regulations. The company specialises in the design and production of high-voltage power supplies and systems for different applications. Power supply specifications cover 300 V to 150 kV, and from mW to kW. Applications range from scientific equipment that demands high precision and performance, through to large telecoms systems that need high reliability and must meet stringent telecoms approvals. In general, the supplies use resonant switched-mode operation. Operating frequencies vary over the range 50 kHz to 500 kHz depending on power and voltage, and harmonics can give emissions in both the conducted and radiated ranges of EMC measurement. In addition, today’s high-voltage power supply control systems utilise processors and FPGAs, which use high-frequency clocks that also contribute to overall emissions. EMC immunity is generally less of a problem, and effects result mainly from the ESD susceptibility of input signals. Table 1 lists the generic and product specifications that the company applies to its product range.
Initially, like many others, Start Spellman’s test engineers decided to evaluate emission characteristics because they expected their products by design to be reasonably immune to electromagnetic effects. Their plan was to use an open area test site (OATS) method in accordance with CISPR 16 — a well known and acknowledged way to test products for emissions. In order to achieve the important accuracy and correlation with conformal test results, the company, in 1995, purchased a Wayne Kerr easy 1 system that includes: • Spectrum analyser Easy 1 system software enables you to select the required test standard from a menu. The system then automatically sets limit lines and lets you make preliminary “quick-look” peak measurements of the radiated emissions. Peaks above the limit are tagged and then evaluated using a quasi-peak detector. The near field probes enable you to pinpoint emission sources from board level components and cables. You can use the LISN for conducted emission tests.
When testing radiated emissions in the open air, the company’s test engineers soon found that ambient RF noise was interfering with the results and accuracy. This type of interference is especially prevalent in the region of 30 to 300 MHz. Having exhaustively examined these emission problems on an OATS, they decided that a screened room would solve the ambient problems. Figure 1 compares ambient RF field strengths in the open and in a screened room. Screened Room Adds Immunity As a further attraction, a screened room on-site allows the company to consider in-house immunity testing. Immunity testing requires the use of a high-power RF source, and the source itself may well interfere with legitimate communications, such as broadcast stations, if signals escape from a screened room environment. Commercial tests for immunity to radiated electromagnetic fields are typically performed from 150 kHz to 1 GHz. A screened room must provide adequate attenuation over at least this frequency range. At low frequency the tests are performed by direct coupling onto cables while at higher frequency the product is subjected to radiated fields. The crossover frequency between these methods varies with the standard used, but radiated tests may extend down to 26 MHz (EN 60601-1-2) while conducted tests may require up to 230 MHz (EN 55104). Both methods give rise to radiated fields that potentially cause interference. Although the RF field is deliberately generated, there is no intention to radiate outside the test environment. Therefore, you can consider a screened room, complete with RF generators as a group 1 product within specification EN 55011. In this case, the emission limit for the RF field generated would be set at 30 dB/mV — at either 10 or 30 metres. This limit implies that you need at least 60 to 90 dB of attenuation to meet compliance. Window Provides Safety The prime task of a screened room in this application is to ensure that a noisy ambient does not interfere with testing. Higher performance screened rooms can achieve 60 to 90 dB attenuation, which is enough to achieve this result. Safety is a major concern for Start Spellman, because personnel have to conduct tests with high voltage within a sealed room. To overcome that concern, the end wall of the room has a 300 x 300 mm screened window, which is RF shielded to maintain the room’s original specification. Regular visual checks take place to ensure personnel working in the room are safe. Doors are clearly important for both RF performance and everyday use with low maintenance. The room has a lightweight steel slam-shut door. The door uses a single row of fingerstock gaskets, which maintain the necessary conductive contact between the opening leaf of the door and the frame of the screened room. Reflections Give No Trouble A screened room provides an RF clean environment that is ideal for accurate conducted emission testing, but Start Spellman’s engineers had concerns about reflections from the room walls during emission testing. To meet their concerns, they use an emission reference source that enables them to quantify the reflection levels and to double check on the test set-up in the room. On balance, the resonant effects of the room have less of an impact on the accuracy of results than the ambient noise levels they experienced on the OATS. Now that Start Spellman can measure emissions, immunity, and ESD with accuracy and repeatability, the company is able to carry out all their development testing in house. The costs (£20k to £25k) associated with the initial purchase of the equipment are relatively small compared to the savings on visits to a formal test house. The company is pleased with the results, which in turn permits self-certification in accordance with EMC regulations. A major benefit is the ability to continually carry out development and testing of new products and variants of existing products to meet new customer requirements. T&ME Graham Muncey is product development manager with Start Spellman, Pulborough, UK. John White is marketing director with Rainford Systems, Wigan, UK, and specialises in EMC chamber design. |



















