Hi All, I can't resist throwing in my 2 cents worth as well. EFT has a pretty steep risetime of current, not quite as fast as ESD but fast enough to cause interesting effects. An EFT generator can produce 100 volts of di/dt across one cm of wire! (a "Faraday" voltage measurement, see http://emcesd.com/tt2002/tt090102.htm for a description of Faraday and Kirchoff voltage measurements). So if a good portion of the generator outptut finds its way to the capacitor being discussed, both the cap leads and any PWB paths associated with it must be very short.
In my experience, EFT problems usually (not always) are fixed by circuit layout, enclosure design, and cables issues rather than protectors or added caps. In general, I consider EFT problems the easiest to fix because the EFT noise is relatively constant (set generator timeout for 5 minutes) which makes it easy to trace the current around circuits with current probes or square magnetic loops. By using phase difference between a reference probe and a moveable one on a dual trace scope can pinpoint not only where currents are flowing, but in which direction as well. ------v Once one can display a problem on the face of some instrument on command, that is 95% of the work in fixing the problem. ------^ Current probe caveat: one can get 1/2 amp of current coupled through parasitic capacitance (a few pf) from a cable with EFT flowing and a current probe body on the cable for a measurement! Imagine what happens in the equipment itself! See my first Technical Tidbit on this at http://emcesd.com/tt061999.htm for more details. In many cases, even voltage measurements are relatively easy to do, just don't use active probes and most unbalanced probes of any type. See http://emcesd.com/tt110100.htm for an easy way to measure voltage drop across on the ground plane of a board with a bent paperclip. The probe (current, voltage, or loop) cable shield must connect 360 degrees to the (hopefully) metal enclosure to prevent conducting noise on the probe cable "antenna" into the system. Just be careful on this. I have had clients that added a voltage probe to the circuit and spent a month debugging the problem it caused by acting like an antenna. My point is you don't have to guess at the problem, EFT can be traced around a circuit and the problem located by simple measurements. Doug drcuthbert wrote: > On the subject of adding caps in circuits for noise immunity. Rather than > just throwing in a value of 0.1 uF, finding it works, and calling it good I > like to use a different approach. Determine the lowest value that will fix > the immunity problem. Then find the largest value that will still allow the > circuit to function properly. Then specify a value somewhere in between; > possibly the geometric mean. In this way you are not using a value "next to > a cliff" that will cause some malfunction in a production run. In a high > volume product just throwing in a value and calling it good is EXTREMELY > poor engineering. > > Dave Cuthbert > Micron Technology > > -----Original Message----- > From: [email protected] [mailto:[email protected]] > Sent: Thursday, January 16, 2003 8:19 PM > To: [email protected] > Subject: Re: I2C bus sensitivity to EFT > > > > I have never worked on design of a product with external I2C bus, but > considering the bus speed of 100 kHz to 3.4 MHz (depending on type) relative > to > about 60-100 MHz BW (depending on definition) of EFT, you may try killing > good > portion of EFT with ceramic caps. > > Be sure that the cable shield is connected on both ends. > > Also, check for the possibility that the EFT may couple to some other > apparently non-critical pin of the IC and then internally cause > susceptibility. > The first suspect in such case would be the reset pin, but often you can be > surprised that other pins may cause problems. I just had a case in which EFT > > would couple to the LED driver on an Ethernet device and cause problems > internally in the chip, leading to packet loss. A cap on the LED driver pin > fixed the problem !! :) > > Neven > >>Hi Forum, >>Has anyone on this forum worked with I2C products and maybe be able to >>advise on the best method to suppress EFT noise that would alow the I2C > > bus, > >>via external cables, to operate as expected e.g. Good EFT devices? How > > best > >>to shield the I2C? Other? >>As always I look forward to your proffesional advice. >>History >>I have a system that connects 3 products (powered from an in-line external >>non-earthed power supply, SELV) using the I2C bus via a 1m shielded and 1m >>shielded curly cable. I use an I2C bi-directional extender IC P82B715 on > > one > >>of the products. The I2C protocol gets corrupted when I appply the >>Electrical Fast Transient (EFT) Test per the EMC Standard EN61000-4-4. >>I have tried shielded cables, several EFT devices and 1nF caps on the > > lines > >>but with little affect. >> >>Kind Regards >>Alex McNeil >>Principal Engineer >>Tel: +44 (0)131 479 8375 >>Fax:+44 (0)131 479 8321 >>email: [email protected] >> -- ___ _ Doug Smith \ / ) P.O. Box 1457 ========= Los Gatos, CA 95031-1457 _ / \ / \ _ TEL/FAX: 408-356-4186/358-3799 / /\ \ ] / /\ \ Mobile: 408-858-4528 | q-----( ) | o | Email: [email protected] \ _ / ] \ _ / Website: http://www.dsmith.org This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/ To cancel your subscription, send mail to: [email protected] with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Ron Pickard: [email protected] Dave Heald: [email protected] For policy questions, send mail to: Richard Nute: [email protected] Jim Bacher: [email protected] All emc-pstc postings are archived and searchable on the web at: http://ieeepstc.mindcruiser.com/ Click on "browse" and then "emc-pstc mailing list"
