Tip on reducing EMI inside and outside the case

Post by **XMEN Gambit** » Tue Jun 18, 2002 5:47 am

EMI inside the case can reduce overclockability and introduce errors into memory or hard drives. EMI outside the case can mess up radio, TV, or other electronic equipment. So here's a few recommendations from AMD. Some are obvious, some aren't. <blockquote>Quote:<hr>  1.The large loop to the power connector is a potential problem. Shorten this length as much as possible by routing the cable in a serpentine manner and tying it with a plastic twist-tie. This solution can reduce emissions by 5 dB. Also, shortening the cable path by routing the fan power cable through the heat sink fins (to allow more direct routing) can decrease EMI emissions.  2.The internal power cable can pick up EMI inside the system and can radiate it through the AC power cord. To avoid this, route the internal power cable next to the metal chassis away from the I/O connectors and as far away from the processor heatsink as possible.  3.Cables inside the system should be routed along the metal chassis and away from EMI sources, such as the microprocessor, clock modules, and high-speed VLSI modules. Power cables for drives should be bundled near the power supply, separate from the ATX power cable, and away from the processor heatsink. Always route the front LED cables away from EMI sources, flat to the chassis, and away from the fan openings. Front USB cables must use a shielded internal cable that is grounded to the chassis at the I/O connector. Generally, route all cables cleanly and keep them away from the memory modules. If there are failing signals at 100-, 300-, 500-, or 700-MHz (100/200MHz memory), or 400, 666, or 933 MHz (133/266 MHz memory), the most likely cause is the DIMMs.  4.The rear plate that touches the rear I/O ports should be made of a metal that has good spring quality, such as stainless steel or spring-hardened steel. Typically, the most vulner-able rear I/O cables are the audio and joystick cables. If EMI emissions drop when these cables are disconnected, then improve the shield-to-chassis grounding for these cables.  5.Most motherboards have a screw connection between the motherboard ground and the chassis, usually within 20 mm to 40 mm of the processor. EMI tests have shown that in some cases insulating, these motherboard ground-points from the chassis ground can reduce EMI emissions. This solution works because some chassis designs offer lower impedance at high frequency than the material (FR4) that the motherboard is made from.  6.The fins on the heatsink may create a waveguide that directs the EMI energy toward the fin ends. If the processor heatsink is suspected of causing EMI problems, replacing it with a heatsink with fins running the opposite direction may reduce EMI levels for that system.  7.In some systems, a ground strap connection to the heatsink can reduce EMI emissions by 4 dB or more. Typically it is better to ground the heat sink to the power supply or to a chassis location close to the power supply.  8.Spread-Spectrum (SS) clocking means the clock signal is intentionally varied to spread the timing clock energy over a small frequency range. Go to the BIOS and make sure this is enabled. Always modulate the spread downward so that the processor never runs above its rated speed.  9.Verify that the chassis is sealed tightly at all seams; even a paper-thin gap is a problem. Remember that it is the length, not the width, of a gap or seam in the chassis that compromises EMI shielding. Empty front drive-bays should have multi-contact EMI shielding covers. Sometimes it is still necessary to add finger-stock material to reduce the length of the gaps between the drive and the chassis when the drive bays are populated. Rivets used on the chassis or power-supply case can also be a problem if they are more than five cen-timeters apart. Too much space between rivets forms a slot antenna. If this condition is suspected, try another brand of power supply with different construction details.<hr></blockquote> I've always wondered what that spread-spectrum clocking was. Anyone know or want to dig up an article? <img height=150 width=115 src=http://www.xmenclan.org/xmengambit.gif border=0 align=left>XMEN member Card-carrying DTM OKL Fish-napper Though a program be but three lines long, someday it will have to be maintained. - The Tao of Programming Life is complex; it consists of real and imaginary parts.

Post by **XMEN Gambit** » Tue Jun 18, 2002 6:33 am

Sigh. Found it myself, though I wasn't looking for it. <img src=http://www.ezboard.com/intl/aenglish/im ... /laugh.gif ALT=":lol"> BIOS optimization guide: <a href="http://www.rojakpot.com/Speed_Demonz/BI ...  BIOS guide: <a href="http://www.lostcircuits.com/advice/bios ...  <img height=150 width=115 src=http://www.xmenclan.org/xmengambit.gif border=0 align=left>XMEN member Card-carrying DTM OKL Fish-napper Though a program be but three lines long, someday it will have to be maintained. - The Tao of Programming Life is complex; it consists of real and imaginary parts.

XMEN_Hammer · Post by **XMEN_Hammer** » Tue Jun 18, 2002 8:51 am

So would this help my TV situation ?????? <img src=http://www.xmenclan.org/avatars/hammer_av.jpg>Proud Member of the XMEN Clan If you run, you will only die tired.

Post by **XMEN Gambit** » Tue Jun 18, 2002 9:20 am

Doubt it. Though you could apply these principles to your computer systems, router, and network wiring as a whole. <img height=150 width=115 src=http://www.xmenclan.org/xmengambit.gif border=0 align=left>XMEN member Card-carrying DTM OKL Fish-napper Though a program be but three lines long, someday it will have to be maintained. - The Tao of Programming Life is complex; it consists of real and imaginary parts.

Tip on reducing EMI inside and outside the case

Tip on reducing EMI inside and outside the case

Re: Tip on reducing EMI inside and outside the case

Re: Tip on reducing EMI inside and outside the case

Re: Tip on reducing EMI inside and outside the case