Once again, I have no religion on isolate or connecting chassis and logic ground but my preference will be multiple connections around the PCB since I believe a lot of PCI cards or disk drives has it shorted internally anyways. I have tested both cases and they don't make any difference to my systems. I also understand why all those classical/conventional single point connection came from as describe below. The purpose is to control the leakage point so that the harmful current will not disrupt the critical circuits operating deep inside the PCB. However, after the ESD induced failure on a server clock board I quoted in previous thread and some observations on newly developed systems, I begin to have second thoughts on these single point contact or isolation rings. The problem with the old school thinking is it is really assuming the ESD discharge can only happen on the peripheral of the PCB. It assumes by isolating the peripheral of the PCB, the discharge current will be forced to flow along the edge of the PCB. This is probably true for old style enclosures with lots of space between the PCB and the enclosure on top and little or no metallic structure at the bottom. This is not the case for all these newer 1U,2U server blades or high wattage PC with Manhattan sky scraper heat sinks or air duck/baffles. The seperation between these metal objects and the enclosure is close enough to become the preferential discharge path if the peripheral of the PCB is completely isolated except for a single point or none. I believe that's what happened in the clock board example I quoted because the board was oriented sideways with the metal can of the oscillator very close to the gasket on the side. In these cases, think 3D instead of 2D and it makes sense to tie the board gnd to the chassis in as many points as you can do just to avoid the zap takes the path through the heat sink of some critical components. As for the BC comment, beside it being speculative, I can also make the top and bottom reference planes to be ground and therefore don't need any plane capacitance to provide the surface current to discharge. -----Original Message----- From: MikonCons@xxxxxxx [mailto:MikonCons@xxxxxxx] Sent: Friday, March 05, 2004 3:25 PM To: si-list@xxxxxxxxxxxxx Subject: [SI-LIST] Re: ESD is a low frequency event -really?? Gentlemen: I apologize for the length of this dissertation, but I hope it may clear up a few seemingly disparate positions opined by others earlier, as well as stimulate some more in-depth thought. As pointed out earlier, ESD is a complex phenomena to model. The results and impacts observed will be radically different depending on the initial charge voltage, rates of closure between the arc points, and total physical structure/configuration of the two participating "bodies" involved. That being acknowledged, about four years ago, I attended an EMC Society presentation (Santa Clara, CA, sorry, don't recall who presented) where multiple, carefully planned and conducted ESD test results were presented. The test items were enclosed in a shielded enclosure to minimize coupling (false) effects on the high-speed oscilloscope probes and cables. The most enlightening findings were that the lower voltages (4k-8k volts) produced rise times in the few picoseconds range (yes, picoseconds). The higher voltages (30k-50k volts) produced much slower rise times (a factor of ~10 as I recall), probably because of the longer and more lossy arc path associated with these voltages. The higher voltage arcs contained significantly more energy (1/2*C*V^2) and were delivered through a longer (higher inductance) and more resistive path; therefore, slower rise times, longer discharge durations, and correspondingly lower frequency content were observed for these strikes before the arc quenched. Bottom line: It is nonsense if one does NOT think that many (lower voltage) ESD events do not contain GHz frequency content with all its negative interference effects. Conversely, very strong low frequency energy content will be observed in higher voltage ESD events. The latter case has particularly nasty magnetic coupling affects. THE FIX: Unfortunately, there is no single magic bullet of design techniques for the wide range of ESD strike characteristics. However, I have achieved considerable success using multiple-layer, via-interconnected (staggered via spacings), chassis ground rings on the periphery of printed circuit boards (PCBs). I have even used sharp points on the outboard edges of the rings to offer "preferred" ESD strike points to protect PCB operational circuits from taking the hit. Some members of the SI List have scoffed at chassis ground rings (which I introduced in 1989 with no prior knowledge that anyone else was using same); therefore, you are free to dissent with me on their value. And clearly, their use in fully enclosed assemblies (as opposed to portable equipment or open enclosure "windows") is not normally necessary as such packaging is not conducive to ESD strike conditions. HOWEVER... My concept for the use of ground rings includes the provision of a low impedance current path isolated from and surrounding the main PCB circuits. This represents a sacrificial intercepting structure that will divert the primary current flow from an ESD strike via the lowest impedance path (always preferred in the natural world) to the main structure, and then to earth. Further, the nature of the surrounding ring tends to divide the strike current so that a 180 degree phase relationship for the resulting strong magnetic fields is created. Circuits that are a few centimeters away experience canceling magnetic fields from the two segments of the ESD current. The circuits on each side of and more near the strike point edge still experience strong fields, but they are reduced (as a minimum) in proportion to the division of strike current flow in the opposite directions along the chassis ground rings; i.e., by at least 6 dB. Other ESD hardening techniques that should be considered in conjunction with the use of chassis ground rings include (a) using buried/stripline traces, (b) NOT routing critical traces along PCB edges (on any layer), and (c) maintaining a continuous signal ground (generally part of a plane) around the PCB periphery (isolated from the chassis ground ring). For superior performance, the ground plane should be tightly coupled to the power plane(s) as well. In my experience on redesigns of deficient PCBs where I incorporated the above noted features, the chassis ground ring structure demonstrated a diversion of at least 95% (typically >=99%) of the ESD energy away from the operational PCB circuits. Item (a) inherently provides shielding generally in excess of 20 dB. Item (b) minimizes any E- and H-field coupling to the trace through spacing. Item (c) will absorb coupled energy in a manner that presents common-mode voltage to the PCB circuits which then tends to be rejected. This result depends on the tight power/ground plane coupling noted earlier, which is why I also strongly urge the use of buried capacitance (BC) power distribution designs. Note also that the magnetic coupling in item (c) will tend to cancel in proportion to the ESD directional/opposite current division in the chassis ground rings; therefore, the coupling mechanism is primarily via capacitively coupled E-field between the rings and the ground plane. Note also that the multiple paralleled chassis ground rings provide a very low impedance (both low inductance and low resistance) to the ESD current, thereby minimizing the E-field strength. All of the above contributes to the robustness of a given design. As a further effect that has not been fully understood (or publicized), the use of BC planes is very beneficial in extending the common-mode voltage coupling characteristic of item (c) out to very high frequencies (GHz region). Hence, BC provides reduced circuit susceptibility to ESD events even into the multi-GHz region. GENERAL COMMENTS: Most PCB designs have unique physical and electrical characteristics; therefore, every Engineer/Designer should think for themselves as to the design risk/environment for ESD and think out the proper mitigation techniques. One should know WHY a specific technique is used based on electromagnetic principles. If one cannot understand the WHY, more education is required, or another vocation is in order. And lastly, please note that many design techniques are supportive to other techniques. If one designs for critical applications (space systems, military, life-threatening, and other man-rated systems as I do), it is better to initially design in extra margins of safety through the use of several complementary techniques, and then (when successful) look to the second generation design for value/cost-engineering modifications. Respectfully, Mike Michael L. Conn Owner/Principal Consultant Mikon Consulting Cell: (408)821-9843 *** Serving Your Needs with Technical Excellence ***. ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List technical documents are available at: http://www.si-list.org List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List technical documents are available at: http://www.si-list.org List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu