Post by thread:[EE] ISM band xtal for lower RFI noise emissions?

I asked here before but my email went to the bit bucket somehow. Do people consider using an ISM band clock xtal in an attempt to move all/most self generated harmonics out of normal radio bands? This is more for keeping radiated noise out of one's own project than for passing RFI emissions conformance testing

On Sun, July 8, 2012 1:02 am, Peter P. wrote: > I asked here before but my email went to the bit bucket somehow. Do people > consider using an ISM band clock xtal in an attempt to move all/most self > generated harmonics out of normal radio bands? This is more for keeping > radiated noise out of one's own project than for passing RFI emissions > conformance testing. I really doubt it would make much difference- your effort is better spent doing good layout- watching where the return current flows, short current paths, etc. Self noise and harmonics are really hard to predict- while *most* are at at a small integral multiple of the fundamental, they can interact, and go all over the place. If you are doing a radio where the fundamental is right on the crystal's frequency, it might make a bit of a difference, but no matter what, good signal integrity will be the key to the best performance. Matt Bennett Just outside of Austin, TX 30.51,-97.91 The views I express are my own, not that of my employer, a large multinational corporation that you are familiar with

I agree, good layout and bypassing IS the answer. We just put the system clock(s) where they need to be, and try not to make noise. We routinely put two layer boards without much ground plane through part 15, Industry Canada, and CE with emissions that are barely detectable, with processors running from 4 - 120 MHz, and a fair amount of peripherals and cables. We don't spend anything extra on shielding, and many of our products are in plastic enclosures. Route power so that both sides go into one side of the bypass cap pads, and out of the other side, go to the chip that you're bypassing. Don't drive signals faster than you need to. You can add series resistors at the driver to slow down the edges, but you'll have to experiment for best compromise. If you can get access to any sort of a spectrum analyzer, and make up a small shielded loop probe, it will help you isolate troublesome points before spending lots of money on testing. RF-Bay amplifiers are a great value for a front end amp. Relatively low cost, and good performance. The number one rule is that it's far easier not to make the noise, than to shield it once you have made it