RE: Explosive Atmosphere Safety Question
You are exactly right. I used to work on automotive fuel electrical systems about 10 years ago. The fuel meter is a current meter, measuring current between 1 to 3 mA (for the one I used). The meter input resistance is greater than 4k ohms. This is the current limit for the fuel sender unit, and the meter is designed to fail in the open circuit mode. The fuel sender does not get enough energy to create a hot enough spark to ignite the gas vapor, since the resistive element is soldered to the sender metal housing, which acts as a heat sink. However, this is not true about the fuel pumps. I have seen the records of several fuel tank explosions due to faulty pumps. On another note, there is a better design for the fuel sender. One company designed a parallel plate device and submerged it in fuel. The fuel acts as the dielectric between the parallel plates. As the fuel level goes down, the capacitance of the parallel plates changes. The device measures the capacitance to determine fuel level. It was said that the device will out last the car, since it has no moving parts to fail. The current on the parallel plates are about 100uA at 500mV, not enough energy to start a fire. I wonder why this device is not used everywhere. George -Original Message- From: owner-emc-p...@majordomo.ieee.org [mailto:owner-emc-p...@majordomo.ieee.org]On Behalf Of Ken Javor Sent: Tuesday, October 08, 2002 5:39 PM To: Price, Ed; 'EMC-PSTC List' Subject: Re: Explosive Atmosphere Safety Question There was a good answer on this subject about there not being an explosive atmosphere within a car fuel tank, but I think there may be another safety factor. If I were designing the sensor system, the meter would be configured as an ohmmeter/ammeter, such that there would be a very high series resistance in series with the tank variable element. The meter empty to full range would actually represent a very small current level. I think it could be designed in such a way that there would be little or no possibility of shorting 12 Volts to the sensor element. I don't recall the specifics, but there is a potential below which resultant sparks have insufficient energy to ignite an explosive atmosphere. I used to know it for JP-4, but it will exist for gasoline as well. on 10/8/02 11:48 AM, Price, Ed at ed.pr...@cubic.com wrote: > > I have recently been pondering the safety design of the typical automobile > gasoline tank fuel quantity sensor (or "sender") assembly. The several > examples that I have seen consist of a float attached to a mechanical pivot > arm. As the fuel level varies, the arm moves, changing the resistance of the > sensing element. > > My first question is how is the mechanical sliding resistive contact > isolated from the explosive fuel/air mixture? Certainly, when the tank is > nearly full, the entire sensor element is submerged in the fuel. But what > happens when the tank is nearly empty, and external air replaces the fuel. > The sensor is then hanging in the fuel/air mixture. In short, how are sparks > avoided at the mechanical sliding sensor contact? > > Secondly, how is the problem of sensor self-heating (during a single-fault > condition) avoided? I can imagine a scenario where the hot vehicle bus is > faulted to the sensor lead. Since a typical sensor element varies between > about 100 Ohms to just a few Ohms, the sensor element could dissipate 25 > Watts or more. This would cause rapid heating of the sensor element, > possibly ending in the burn-out of the resistive element (inside the fuel > tank). Again, how is this ignition scenario prevented? > > I am hoping to get a few informed comments before I go out and buy one for a > sacrificial dissection. (I think I became an engineer because I never could > get those frogs to work again after re-assembly.) > > > Regards, > > Ed > > > Ed Price > ed.pr...@cubic.com > Electromagnetic Compatibility Lab > Cubic Defense Systems > San Diego, CA USA > 858-505-2780 (Voice) > 858-505-1583 (Fax) > Military & Avionics EMC Is Our Specialty > Shake-Bake-Shock - Metrology - Reliability Analysis > > > --- > 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: > majord...@ieee.org > with the single line: > unsubscribe emc-pstc > > For help, send mail to the list administrators: > Ron Pickard: emc-p...@hypercom.com > Dave Heald: davehe...@attbi.com > > For policy questions, send mail to: > Richard Nute: ri...@ieee.org > Jim Bacher: j.bac...@ieee.org > > 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" > -- Ken Javor EMC Compliance Huntsville, Alabama 256/650-5261
RE: Shielding Effectivness Question
RE: Product Safety: A Matter of Law or Litigation?Look for a book on RF or microwave. My favorite is Fields And Waves In Communication Electronics. Look in the section on "waveguides." Operate the waveguide below "cutoff." A small aperture in a sheet metal with finite thickness is essentially a very short waveguide. Calculate the attenuation of a waveguide operating below cutoff. The radiated emission from one aperture may seem insignificant, but when you have 100 apertures radiating in phase may cause you to fail FCC A. George Tang -Original Message- From: owner-emc-p...@majordomo.ieee.org [mailto:owner-emc-p...@majordomo.ieee.org]On Behalf Of Neven Pischl Sent: Tuesday, June 05, 2001 9:16 AM To: emc-p...@majordomo.ieee.org; si-l...@silab.eng.sun.com Subject: Shielding Effectivness Question I would appreciate if anyone could let me know if there are any references (books, application notes, anythig ..) that deal with shielding efectivness in cases when a source is close to an (electrically small) opening in a shield (enclosure). In such a situation, the field will penetrate through the hole and leak even if the size is much smaller than the wavelength. I am particularly interested in situation when high-frequency source, such as a PCB edge or a component operating at (say) 1 GHz and above is in proximity of the venting holes, "small" gaps in the chassis etc. All references that I have deal with uniform plane wave propagating incident to a metal plane with a slot or hole, in which case it is enought o have electrically small size of the opening (e.g. lambda/10) to efficiently block any field propagation through the barrier. I can't find any useful reference that deals in any analytical way with the situation I am intersted in. I believe I might get some answers using some of the simulation programs, but at the moment I am more intersted in the analysis of the problem than in simulating it. Thank you, Neven Pischl
Re: Conductive Coating
This will work only if the contact resistance between the probe and the plated surface is zero. If the contact resistance dominates, the reading you see will be the same whether the probes are spaced one inch or two inches. That was what I saw when I used regular probes. I used copper foils for that reason. To go one step further, you can assume the contact resistance as X and the surface resistance as Y. You can take a few readings and solve for X and Y. :-) Ken Javor wrote: > With regards to the question at end of Macy's e-mail. Count the number of > squares between probe points. Say the probes are 2" apart. If the surface > containing the two points is at least 2" wide, then it only takes one > square, and the resistance measured is numerically equal to the ohms/square. > If the width of the surface were but 1", then there would be two squares, > and the measured resistance would be twice the ohms/square of the material. > > -- > >From: "Robert Macy" > >To: "George Tang" , "Westerdahl, Eric" > > >Cc: "'EMI-PS Group'" > >Subject: Re: Conductive Coating > >Date: Wed, Nov 3, 1999, 10:18 PM > > > > > > > A small pedantic point here: > > > > Coatings are usually rated in "ohms per square" which is dimensionless. > > > > Looking at the resistance formula: > > > > R = rho * length / Area = rho * length / ( width * thickness ) > > > > Given fixed rho and fixed thickness (of the coating) you can see that you're > > left with a dimensionless: > > > > R = K * length/width > > > > If the length and the width are the same, you have a square. The resistance > > of a square of any magnitude will be the same. > > > > Now, to relate Ohms per square to the reading one gets when one sticks two > > probes down on it? > > > >- Robert - > > > > > > -Original Message- > > From: George Tang > > To: Westerdahl, Eric > > Cc: 'EMI-PS Group' > > Date: Wednesday, November 03, 1999 10:17 PM > > Subject: Re: Conductive Coating > > > > > >> > >>There are many different types of conductive coatings available. Silver > > paint > >>is > >>very conductive, less than 5 ohms per square inch. But it is not as > > scratch > >>resistant > >>as sheet metal surfaces. Electroless copper / nickel plating is very > >>conductive and > >>durable. You can get as low as 1 ohm per square inch. If you don't, your > >>plating > >>is not thick enough. This plating should pass the safety fault current > > test, > >>as long as > >>the safety ground wire makes "surface area" contact with the plating and > > not > >>"point" > >>contact. This plating has 60 dB shielding effectiveness for frequencies > > above > >>30 MHz, > >>since it is much thicker than the skin depth. You need the thickness for > > the > >>safety fault > >>current. The best feature of the plating is that it allows you to mold > > your > >>chassis into > >>one piece of plastic with no extra metal pieces to assemble. It's kinda > > nice > >>that way. :-) > >> > >>go to www.ccoatings.com > >> > >>or call (972) 851-0460 > >> > >> > >> > >>George Tang > >> > >> > >> > >>"Westerdahl, Eric" wrote: > >> > >>> Our company has decided to use a conductive coating to mitigate some EMI > >>> problems on one of our units. We have not used this method before. I > > have > >>> a question as to the correct resistivity of the coating. What range > > should > >>> I be looking at, and does the range change if the frequency of the strong > >>> signal are high or low? > >>> > >>> The equipment is IEC 950 and EMC Directive stuff with many noisy DC motor > >>> and motor controller combinations. Most of the signals we are concerned > >>> about are at the lower end of the CISPR 22 region. 30 to 150 MHz. > >>> > >>> Eric Westerdahl > >>> Regulatory Engineer > >>> Roll Systems, Inc. > >>> eric_westerd...@rollsys.com > >>> > >>> - > >>> This message is coming from the emc-pstc discussion list. > >>> To cancel your subscription,
Re: Conductive Coating
Here is one way to make a rough measurement: Take a coated surface. Take a knife and carve a 3" x 1" plated area on the surface. Use a DMM to verify that the isolated rectangular surface is electrically separate from the rest of the plating. Cut 2 pieces of copper foil into 1" x 1" square. Solder the square copper foils to the 2 lead wires on the DMM. Now press the 2 square foils onto each end of the 3" x 1" plated area firmly with the palm of your hand (with 3.14 lb of force exactly :-) ) so the current flows through the 1" x 1" plated surface in the middle of the 3" x 1" area. Take a reading on the DMM. It would probably say 1.3 ohms. Press the 2 copper foils onto each other firmly with the same force and take a reading. It should say about 0.4 ohms. Assuming your hand is a poor conductor and has minimal effects on the readings, the resistance of the plating is about 1 ohm per "square inch." You really should use a micro-ohm meter to measure this to get accurate readings. But accurate readings are not terribly important to me as long as they are less than 1 ohm. George Robert Macy wrote: > A small pedantic point here: > > Coatings are usually rated in "ohms per square" which is dimensionless. > > Looking at the resistance formula: > > R = rho * length / Area = rho * length / ( width * thickness ) > > Given fixed rho and fixed thickness (of the coating) you can see that you're > left with a dimensionless: > > R = K * length/width > > If the length and the width are the same, you have a square. The resistance > of a square of any magnitude will be the same. > > Now, to relate Ohms per square to the reading one gets when one sticks two > probes down on it? > >- Robert - > > -Original Message- > From: George Tang > To: Westerdahl, Eric > Cc: 'EMI-PS Group' > Date: Wednesday, November 03, 1999 10:17 PM > Subject: Re: Conductive Coating > > > > >There are many different types of conductive coatings available. Silver > paint > >is > >very conductive, less than 5 ohms per square inch. But it is not as > scratch > >resistant > >as sheet metal surfaces. Electroless copper / nickel plating is very > >conductive and > >durable. You can get as low as 1 ohm per square inch. If you don't, your > >plating > >is not thick enough. This plating should pass the safety fault current > test, > >as long as > >the safety ground wire makes "surface area" contact with the plating and > not > >"point" > >contact. This plating has 60 dB shielding effectiveness for frequencies > above > >30 MHz, > >since it is much thicker than the skin depth. You need the thickness for > the > >safety fault > >current. The best feature of the plating is that it allows you to mold > your > >chassis into > >one piece of plastic with no extra metal pieces to assemble. It's kinda > nice > >that way. :-) > > > >go to www.ccoatings.com > > > >or call (972) 851-0460 > > > > > > > >George Tang > > > > > > > >"Westerdahl, Eric" wrote: > > > >> Our company has decided to use a conductive coating to mitigate some EMI > >> problems on one of our units. We have not used this method before. I > have > >> a question as to the correct resistivity of the coating. What range > should > >> I be looking at, and does the range change if the frequency of the strong > >> signal are high or low? > >> > >> The equipment is IEC 950 and EMC Directive stuff with many noisy DC motor > >> and motor controller combinations. Most of the signals we are concerned > >> about are at the lower end of the CISPR 22 region. 30 to 150 MHz. > >> > >> Eric Westerdahl > >> Regulatory Engineer > >> Roll Systems, Inc. > >> eric_westerd...@rollsys.com > >> > >> - > >> This message is coming from the emc-pstc discussion list. > >> To cancel your subscription, send mail to majord...@ieee.org > >> with the single line: "unsubscribe emc-pstc" (without the > >> quotes). For help, send mail to ed.pr...@cubic.com, > >> jim_bac...@monarch.com, ri...@sdd.hp.com, or > >> roger.volgst...@compaq.com (the list administrators). > > > > > >- > >This message is coming from the emc-pstc discussion list. > >To cancel your subscription, send mail to majord...@ieee.org > >with the single line: "unsubscribe emc-pstc" (without the > >quotes). For help, send mail to ed.pr...@cubic.com, > >jim_bac...@monarch.com, ri...@sdd.hp.com, or > >roger.volgst...@compaq.com (the list administrators). > > > > - This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to majord...@ieee.org with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to ed.pr...@cubic.com, jim_bac...@monarch.com, ri...@sdd.hp.com, or roger.volgst...@compaq.com (the list administrators).
Re: Conductive Coating
There are many different types of conductive coatings available. Silver paint is very conductive, less than 5 ohms per square inch. But it is not as scratch resistant as sheet metal surfaces. Electroless copper / nickel plating is very conductive and durable. You can get as low as 1 ohm per square inch. If you don't, your plating is not thick enough. This plating should pass the safety fault current test, as long as the safety ground wire makes "surface area" contact with the plating and not "point" contact. This plating has 60 dB shielding effectiveness for frequencies above 30 MHz, since it is much thicker than the skin depth. You need the thickness for the safety fault current. The best feature of the plating is that it allows you to mold your chassis into one piece of plastic with no extra metal pieces to assemble. It's kinda nice that way. :-) go to www.ccoatings.com or call (972) 851-0460 George Tang "Westerdahl, Eric" wrote: > Our company has decided to use a conductive coating to mitigate some EMI > problems on one of our units. We have not used this method before. I have > a question as to the correct resistivity of the coating. What range should > I be looking at, and does the range change if the frequency of the strong > signal are high or low? > > The equipment is IEC 950 and EMC Directive stuff with many noisy DC motor > and motor controller combinations. Most of the signals we are concerned > about are at the lower end of the CISPR 22 region. 30 to 150 MHz. > > Eric Westerdahl > Regulatory Engineer > Roll Systems, Inc. > eric_westerd...@rollsys.com > > - > This message is coming from the emc-pstc discussion list. > To cancel your subscription, send mail to majord...@ieee.org > with the single line: "unsubscribe emc-pstc" (without the > quotes). For help, send mail to ed.pr...@cubic.com, > jim_bac...@monarch.com, ri...@sdd.hp.com, or > roger.volgst...@compaq.com (the list administrators). - This message is coming from the emc-pstc discussion list. To cancel your subscription, send mail to majord...@ieee.org with the single line: "unsubscribe emc-pstc" (without the quotes). For help, send mail to ed.pr...@cubic.com, jim_bac...@monarch.com, ri...@sdd.hp.com, or roger.volgst...@compaq.com (the list administrators).
Re: explanation of ESD events with coins in baggie.
Barry, Your question is very much justified. I use the word "ESD" because in my mind, I play the image of the striking screw driver in slow motion. If you imagine the nickel surface robs the screw driver of its electrons to charge up the screw driver, then the electric field builds up as the screw driver moves away from the nickel surface to the point to cause air break down and screw driver discharge. In slow motion, this is exactly the ESD process. But in real-time, this is RF. Different perspective makes worlds of differences. :) Thanks, George Tang (Bailin Ma) wrote: > George, > > I think your reasoning is convincing to me. Thanks. > But please allow me to pose a silly question: Are we really sure those > kinds of "shaking coins" interferences can be categorized as ESD > (Electrostatic Discharge) problem? > In your experience, for instance, the screw driver did not carry any > electrostatic charge before striking the nickel surface. Right? We don't > know if coins, keys, and metal door got electrostatic charged before > jangling or slamming in Doug Mckean's experiences, either. > ASSUMING friction and striking between different materials would cause fast > electron transition between materials, back and forth, and then produce > strong RF EM waves, there is no Electrostatics involved. > If we are not sure whether or not those objects were electrostatic charged > before friction, can we try to see if it makes difference? > > Best Regards, > Barry Ma > > > - > Original Text > From: "George Tang" , on 3/15/99 4:25 PM: > Barry, > > No, I did not measure the 15kv on the scope, but I suspect that was the > case. > Here is my reasoning: The system was well shielded with aperture size > smaller > than 1/2 inch with few apertures. The system passed FCC B emi limits with > 8 dB > margin. It seemed very unlikely for a power plane in such a system to pick > up > 8 volts of radiated noise, since large power planes are not efficient > antennas. And to charge the power and ground planes with many caps up to 8 > volts at 100 MHz fundamental and GHz harmonics would seem to require very > high > power of radiated energy, assuming the lumped circuit model is used for the > caps and planes at 100 MHz. But on the other hand, the boards are grounded > to > the chassis, so if the chassis had ESD noise of many kv conducted to it, > it > could generate 8 volts on the power to ground plane. The chassis is low > impedance, so high current noise is likely. With radiated noise, even at > near > field, the propagation impedance would still be higher than conducted, so > high > current noise seemed more unlikely. The system had already passed 15kv ESD > air > (and accidental contact discharge) on most all parts on the chassis, then > the > screw driver noise maybe higher than 15 kv conducted ESD. You can see > there > are many assumptions used. But trying to measure the screw driver to > chassis > ESD voltages directly with differential probes is difficult because a large > current loop formed by the probes is required due to the moving parts and > the > probes may be damaged by the ESD. I did not measure the screw driver > voltage, > but I guessed that it was high voltage/current conducted ESD that caused > the > system error. > > Hope that helps. > > George Tang > > (Bailin Ma) wrote: > > > George, > > > > You mentioned: "the ESD generated by the metal surface was much higher > than > > 15kv." at the end of your note. > > Does that mean you measured the noise between power and ground planes > using > > digital scope, and got the higher than 15 KV reading on the scope when > > striking a screw driver against the nickel plated surface? > > > > Thank you. > > Barry Ma > > Anritsu Company > > Morgan Hill, CA 95037 > > > > - > > Original Text > > From: "George Tang" , on 3/15/99 1:23 PM: > > Douglas, > > > > I have seen similar events in a different way. Years ago, I helped > design > > an electronic system using plastic chassis with nickel surface plating. > > The > > system passed 15kv ESD air discharge and 8kv contact. But in the > hardware > > lab, the system gets data error everytime a piece of metal (like a screw > > driver) is striked against the nickel plated surface on chassis. A > digital > > scope is used to measure the noise generated on power and ground planes > on > > the PCB inside the chassis and the scope captured a noise voltage as high > > as > > 8 volts peak to peak on the PCB from a few hundred MHz to beyond GHz.
Re: explanation of ESD events with coins in baggie.
Douglas, I have seen similar events in a different way. Years ago, I helped design an electronic system using plastic chassis with nickel surface plating. The system passed 15kv ESD air discharge and 8kv contact. But in the hardware lab, the system gets data error everytime a piece of metal (like a screw driver) is striked against the nickel plated surface on chassis. A digital scope is used to measure the noise generated on power and ground planes on the PCB inside the chassis and the scope captured a noise voltage as high as 8 volts peak to peak on the PCB from a few hundred MHz to beyond GHz. The PCB was very well decoupled with power next to ground planes and many on board capacitors. This puzzled me at first. But I remembered a very knowledgeable mechanical engineer once told me to never use nickel material in an application where friction takes place. Nickel has a very hard and rough surface, so in a frictional application, it always damages the mating surface. Maybe this explains the events that you saw, and the ESD generated by the metal surface was much higher than 15kv. Regards, George Tang Douglas McKean wrote: > Hans, > > That's certainly an interesting explanation, but does > not correlate to at least three different scenarios. > > 1) A calibrated ESD simulator in self discharge >mode at 15KV. When the results of the ESD >simulator are compared to the results of the >coins, the coins have a fairly wideband constant >level from 0 - 2 GHz. Both start off at roughly >the same level with the only the coins remaining >constant throughout. The ESD simulator has approx >a -20dB per octave drop off. > >A side interest is that on the display of the SA has >an IF overload indication. This tells me that the >transients from the coins are quite possibly a lot >higher and much quicker than what the SA can handle >within the sampling window. > > 2) The level from the coins is proportional to the >dissimilarity of the metals of the coins. A bag of >quarters has a lower profile than a bag of quarters >and pennies. Thus, there is some function due to >electronegativity differentials. Actually, a >significant amount of difference. > > 3) I can cause the same effect by sliding the coins >back and forth as a group within the bag. Thus, >the coins are in at least incidental contact with >each other so that differing potentials amongst >the coins is minor. > > I'm not sure if anyone knows the reason. > > Regards, Doug McKean > > At 11:11 AM 3/11/99 -0800, Hans Mellberg wrote: > > > > > >There is an expanation for this seemingly unlikely event. > > > >Having coins in a baggie and jingeling them causes the following > >events to occur: > > > >The rubbing of a coin against the polymer causes triboelectric > >charging of both the coin and localized areas of the bag. Since there > >are multiple coins, each coin will charge at some voltage level but > >not necessarily the same as another coin. When two coins of different > >charged voltages come within dielectric breakdown distances, a > >discharge will occur from one coin to the other in order to equalize > >the charge distribution (q1=C1V1 and q2=C2V2. When they touch, the new > >q1 will be C1V3 and q2= C2V3 where V3=(q1+q2)/(C1+C2)). Since coins > >are electrically small with very small capacitances, the expected > >discharge waveform has a very fast risetime hence the radiation at the > >GHz region. There will also be discharges from the localized charged > >areas of the polymer to coins of different voltages. While separating > >two charged surfaces from each other, the voltage rises significantly > >since the capacitance is being reduced and the conservation of charge > >must be preserved which is the basis for tribolectric voltage > >generation. > >Hope that helps > >Hans Mellberg > > > > > >---b...@namg.us.anritsu.com wrote: > >> > >> Hi Douglas, > >> > >> What you described is very interesting! But I cannot understand > >"Jingling > >> change in a ziplock bag produces very high levels of super fast > >transients > >> up into the GHz range." It seems to me that jingling coins, jangling > >keys, > >> and slamming metal door would certainly produce acoustic waves. How > >come > >> they also produced electromagnetic waves? If do, under what > >conditions? > >> What is the mechanism to produce "very high level" of transient EM > >waves? > >> Did that company incorporate those kinds of "Jingling change in a > >ziploc
