--- On Thu, 26 Mar 1998 13:14:35 -0500 Scott Douglas <[email protected]> wrote: > Hello Group - > > Here's a problem for you all to think about and hopefully help solve. > > I have a room in the corner of my building where I perform EMI pre-scans. > Occasionally throughout the day I have a major interference show up on my > analyzer. It does not come from the EUT. This is broadband, from 30 MHz to > 1 GHz, at levels up to 70 or 80 dBuv. It is not like a flat carrier but > rather more like PWM driven motor noise. It lasts from 15 seconds to 1 > minute in duration. Then ambient returns to normal. > > We are a typical ITE manufacture. Our Genrad tester, wave solder, air > compressors, motor generators, air conditioners, plumbing waste pumps do > not appear to be the cause although none have been absolutely ruled out. > The frequency and duration would seem to eliminate them however. > > We have swamp on one side,
Not a likely source ;) > the interstate highway on another, could be a very noisy vehicle ignition system which is driven past your location several times a day; do you see any schedule to the noise? (If you listen to ignition noise, you can actually hear the engine RPM, and you can tell an automatic from a manual transmission car. You can hear a vehicle as it slows for a curve, etc. Didn't Yogi Berra say that you can hear a lot when you listen?) > fields on the third again, not likely > and the town on the last side. Any complaints about EMI in the town? The noise spectra you are describing would create many gripes in an urban environment. Indeed, anything broadband enough to cover 30 MHz to 1 GHz is also likely has energy down to 1 MHz. This suggests you might be able to use a cheap loopstick antenna AM radio as a sniffing tool. The loopstick is quite directional. Practice doing a quick bearing sighting on a known location AM radio station. Then, be ready with the radio when the noise strikes again. > The only potential source I can > find is a medical clinic about 350 meters away from us. They are the > nearest building (the next being twice that far) and have an X-ray > machine. Yesterday I went there at lunch and asked was the X-ray in use > that day and they said yes but could not tell me a specific time that it > was used. A reasonable possibility. > A software engineer (Uh Ohhh; these are guys who don't believe in harmonics and parasitics!) > here tells me that he has experienced radio problems > in the parking lot outside my corner of the building. When he parks in a > certain spot, his FM radio (the entire band) does he ever listen to AM? What about that band? > gets wiped out by a > zhhzhhzhhh kind of noise. And this observation is not time dependent? You can repeat this at any time? This would sound more like a multipath wave cancellation node; the direct signal plus a reflection cancells the field strength at an unlucky location. Move a few feet, and the cancellation goes away. The signal magically "emerges from the noise." This may be a distraction from your problem. > Move the car several feet any direction the > problem goes away. This closely resembles what I see on the analyzer. And just what do you "see" on the analyzer? Maybe you could set the analyzer up to a normally quiet freq, set the span to zero, trigger on video, and capture the time domain view of the signal. Is it random, or do you see pulse modulation? Does the PRR suggest anything? > Our > analog guru suggests it is not the x-ray system as x-rays only last a few > seconds. A one minute x-ray would burn you up he says. Yes, the actual X-ray emission. But maybe the HV power supply runs for a minute before the tube is gated on. You may be seeing the spectra of an HV insulation breakdown. > I never noticed this interference in another corner room of my building > diagonally opposite my present location nor have I noticed the FM radio > problem myself. > Questions: > 1. What are the possible sources of such interference? Still many; more data is needed. Work with that clinic; at least you should try to rule that very likely source. > 2. Is the problem radiated or conducted? I suspect radiated as the signal > goes away when you disconnect the antenna from the analyzer. About all that eliminates is an internal spurious in the SA. The path could be totally radiated, from far away directly to your test antenna. It could come into your building on the powerlines, only to be radiated locally. I could be generated locally, with radiated AND conducted components. > 4. Why does this problem show up in only a small specific location on our > property? You need more data to substantiate this. > 4. What are the potential solutions? Find the source. Fix the source (if the source owner will let you). Break the path. Work 3rd shift. Move out of Dodge. Live with it. Fix it in software with a -60dB correction factor. Retire. Get promoted. > I will be happy to answer any other questions you think of (if I can) and > I look forward to the interesting comments I will receive from you all. > Thanks in advance. > Scott Douglas > [email protected] ---------------End of Original Message----------------- Scott: Tracking down a noise burst that lasts 15 to 60 seconds and occurrs randomly every couple of hours is not easy. You will need more observations of the condition, plus a bit of luck. Let us know how things turn out. Ed -------------------------- Ed Price [email protected] Electromagnetic Compatibility Lab Cubic Defense Systems San Diego, CA. USA 619-505-2780 List-Post: [email protected] Date: 03/26/98 Time: 12:27:01 --------------------------
