Chris & other, Thanks for the reply. Chris has given a great alternate method. I would like further add on to this test method/approach.
Those second source components (ferrite or oscillator) that we get of course have to meet the PCB footprint as well as meeting the primary specification. For ferrite, let's now say the second source component has the correct impedance at 100MHz and current rating. However, the impedance curve of ferrite from different source are mostly of different response curve. It would be difficult to judge whether does it degrades the final product emission level. In view of this, I might add that we can make use of the final product scanning results for a guide to make a final judgement. For example, if the final product has operating frequencies of 34MHz, the worst three case frequencies are 68MHz, 102MHz & 136MHz with passing margin of 4dB, 2dB & 6dB margin. Comparison of the ferrite bead impedance at these frequencies yield the following reading, Freq 1st source 2nd source 68 67 ohm 50 ohm 102 98 ohm 96 ohm 136 80 ohm 85 ohm How can we have a simple calculation to state that this 2nd ferrite is OK on the product? I would like to hear your view of such formula dB change = 20 log (1st impedance) - 20 log (2nd impedance) At 68MHz, dB change = 20 log 67 - 20 log 50 = 2.54 dB With this 2.54dB, the final product is still passing with 4-2.54 = 1.45dB. Same for 102MHz yield 1.82dB and 136MHz yield 6.53dB passing. Would this be the worst case scenario that will occur? How's the group view on such approach? As for crystal/oscillator, I don't have any idea on paper approach except to perform some in-house 1m scanning. Anyone that has one, I would greately like to hear from him/her. On using PCB EM scanning system for component qualificaiton, I have some question. How will we know that the reading obtained are ensuring that the final product is still meeting the CISPR 22 limits? What margin should be given when comparing the PCB with 1st and 2nd source component? Will we overkill the product due to this set margin, due to no correlationship between near field & far field signal strength? Thanks again for the response. Looking forward to your response. Regards Koh "Maxwell, Chris" wrote: > Wait a minute! > > Buying a board scanning system to evaluate different vendors for ferrites > and oscillators? My company doesn't have that kind of money to throw > around. These systems can cost 10's of thousands of dollars. Ferrite > beads cost about a dime for a dump truck load. If it costs $10,000 to > evaluate a second source for ferrites, I 'll stick with the ferrites I have. > I'd like to offer a lower cost alternative. > > Koh Nai asked about what specifications were important for qualifying > alternate sources for ferrites and oscillators. > > When it comes to ferrites, I look at three things: I look at the PCB > footprint (it won't work if it won't fit). I look at the impedance curve > and the current capacity. If all three of these specifications are equal or > better than what I need, I accept them. I don't even consider re-testing > for emissions if I have checked these three specifications. Alternate > sources for ferrites can be qualified for the cost of reading a spec sheet. > > Oscillators are a different story. When one of my digital design colleagues > wants to change oscillators. They consider its PCB footprint, the output > frequency, its ambient stability, its temperature stability and its load > driving capability. If it meets their needs; then I take a circuitboard > with the existing oscillator and run a near field probe over it near the > oscillator until if find a location of maximum near field emissions. (If > you really are strapped for cash, you can make your own near field probe.) > I write down the exact position and orientation of the near field probe and > I either print out or write down the spectrum analyzer readings. I then put > the new oscillator on the exact same board and repeat the experiment. If > the measurements are close (within a dB or two) I don't worry. If the > measurements are more than 4 dB higher, then I look further. Then I > consider: testing the whole unit with the new oscillator with my antenna > set up 1 meter away in-house, or re-testing the unit for emissions at an > OATS, or not using the new oscillator. > > ONE WARNING: if the new oscillator is at a different frequency, then the > method above WILL NOT yield any useful results. > > One thing that we have done with new designs is to put a 1206 surface mount > PCB footprint in line with the oscillator output. We start our testing with > a 0 ohm resistor. If we run into problems, we can put either a ferrite bead > or higher value resistor in this position to "cool off" the oscillator. > This has worked well with oscillators under 100MHz. I don't know if it will > work for faster oscillator. > > I know that there are problems with using near field probes to make such > correlations, however using a board scanning device would cost much more > than a near field probe and still only be measuring near fields. Even so, > if I had the budget, I'd love to try one out. > > To me, the real method of doing this starts with the initial testing of your > product. I try to get more than a 5dB margin during the initial testing. > With these margins, I don't need to worry so much about component > differences. I know that this is sometimes not possible. I have sacrificed > margins in order to get a product's testing done and released (I don't get > paid if we don't ship.) The problem is, without margins, I need to worry > more about component differences. > > Another point to remember is that EN 55022 and other emissions standards go > by the "80 %" rule. A product "passes" if we are confident that 80% of > the units that we ship meet the emissions requirements. Anybody who wants > to dispute whether your units pass or not is REQUIRED to test up to 7 > samples in order to get enough data to use statistical methods to compute > the confidence factor. One failing unit does not equal a guilty verdict. > (Of course, if one unit is failing by 20dB, that's a problem.) The people > at CISPR put this slack into the limits to allow for slight component > differences and measurement uncertainty. They allowed us the slack, but it > is our responsibility to use it with common sense. > > These are my two cents worth and definitely don't reflect the opinions of my > employer. (Why would they need me if I thought the same way they do?) > > Have a great day! > > Chris Maxwell, Design Engineer > GN Nettest Optical Division > 6 Rhoads Drive, Building 4 > Utica, NY 13502 > PH: 315-797-4449 > FAX: 315-797-8024 > EMAIL: [email protected] > > > -----Original Message----- > > From: Ralph Cameron [SMTP:[email protected]] > > Sent: Wednesday, September 20, 2000 11:16 PM > > To: Tony J. O'Hara; Koh Nai Ghee > > Cc: EMC-PSTC > > Subject: Re: Component Qualification > > > > > > Tony: > > > > I have sold these systems and also used them and they are good for > > graphing > > a profile of a scanned board and giving the field intensity vs postion > > over > > the board area. You can quickly find trouble spots using the computer > > interface. > > > > Ralph Cameron > > > > ----- Original Message ----- > > From: "Tony J. O'Hara" <[email protected]> > > To: "Koh Nai Ghee" <[email protected]> > > Cc: "EMC-PSTC" <[email protected]> > > Sent: Wednesday, September 20, 2000 4:42 PM > > Subject: Re: Component Qualification > > > > > > > > > > You may want to look at using a PCB Electromagnetic Scanning System! > > One > > > of the advertised uses for these test devices is for quickly comparing > > EMC > > > performance when component changes are made etc.! I believe there are 4 > > > different manufacturers who make these devices. The one that I'm just > > > starting to learn about is made by EMSCAN in Canada. Their web is > > > www.emscan.com > > > Maybe someone who has & uses one of these scanners can provide an > > > experienced viewpoint? > > > Regards > > > Tony > > > Colorado > > > > > > ------------------------------------------- > > > This message is from the IEEE EMC Society Product Safety > > > Technical Committee emc-pstc discussion list. > > > > > > To cancel your subscription, send mail to: > > > [email protected] > > > with the single line: > > > unsubscribe emc-pstc > > > > > > For help, send mail to the list administrators: > > > Jim Bacher: [email protected] > > > Michael Garretson: [email protected] > > > > > > For policy questions, send mail to: > > > Richard Nute: [email protected] > > > > > > > > > > > > > > > ------------------------------------------- > > This message is from the IEEE EMC Society Product Safety > > Technical Committee emc-pstc discussion list. > > > > To cancel your subscription, send mail to: > > [email protected] > > with the single line: > > unsubscribe emc-pstc > > > > For help, send mail to the list administrators: > > Jim Bacher: [email protected] > > Michael Garretson: [email protected] > > > > For policy questions, send mail to: > > Richard Nute: [email protected] > > > > ------------------------------------------- > This message is from the IEEE EMC Society Product Safety > Technical Committee emc-pstc discussion list. > > To cancel your subscription, send mail to: > [email protected] > with the single line: > unsubscribe emc-pstc > > For help, send mail to the list administrators: > Jim Bacher: [email protected] > Michael Garretson: [email protected] > > For policy questions, send mail to: > Richard Nute: [email protected] ------------------------------------------- This message is from the IEEE EMC Society Product Safety Technical Committee emc-pstc discussion list. To cancel your subscription, send mail to: [email protected] with the single line: unsubscribe emc-pstc For help, send mail to the list administrators: Jim Bacher: [email protected] Michael Garretson: [email protected] For policy questions, send mail to: Richard Nute: [email protected]
