Ceramic resonator and rf antenna proximity

[jeremyabel]

Hello!

So I'm laying out a board here, components include a bluetooth module w/ antenna, and a 7.37MHz ceramic resonator . Now, this is actually the 2nd revision of this board, and I need to add said ceramic resonator, which I know needs to be within reasonably-close proximity to the microcontroller I have it hooked up to. The design is pretty crowded for space though, and the only location I can find is literally right next to the bluetooth module, near the antenna. If you look at the picture in the sparkfun link up there, the resonator would be going about 1.5mm, next to where the R in RN-42 is.

I'm no RF guy or anything, but would having a crystal oscillator that close to an RF antenna be a bad idea? It kind of seems like the sort of thing that might be a bad idea. Is this something I need to keep in mind? Cause I can certainly move stuff around to find a better place for it if it's something I should be concerned about. If it helps, the crystal is for USART timing for communication from the microcontroller to the bluetooth module.

[rs20]

I'm prepared to yield to other, more experienced people, but my instinct is that this will probably be fine because, among other things, 7.37 MHz is several orders of magnitude away from the ~2.4GHz that Bluetooth uses. That means that while the crystal lazily swings up and down once, your bluetooth has emitted/received over 300 pulses -- the crystal is basically DC as far as the bluetooth module is concerned.

[T3sl4co1l]

You mean beside the shielded part of the module?

Sure, no problem.  In fact that should all be ground plane surrounding that thing, anyway.  (You put ground plane / pour on your boards, right?...)

Keep everything away from the antenna end, especially signal traces and ground pour.  You usually want a keepout on the board in that region.  The datasheet probably talks about this.

If you're concerned about HF coupling into things, you can add a few extra pF to most any signal trace without too much worry.  The resonator probably needs some explicitly, or already is a capacitor (among other things).

If you aren't concerned about HF coupling into things, consider that most CMOS inputs are high impedance (especially around low power clock oscillators!), and you can end up with weird glitches and extra transitions and mysterious DC offsets or transient weirdness (since BT isn't continuous wave after all, but goes on and off!).

Bipolar amplifiers are especially prone to detecting (rectifying and making a DC offset from) signals well beyond their bandwidth -- for example, a little 100kHz ripple will already cause an LM358 to behave slightly odd (not quite correct at DC), even though that's still well within its GBW; interference in the MHz+ range is actively detected and rectified by the front end transistors (which are just diodes, first and foremost).  Not sure if it will do anything with 2.45GHz, but GSM cell phones are notorious for pumping noise into everything, and that's 900MHz+.

Tim

[jeremyabel]

You mean beside the shielded part of the module?

Sure, no problem.  In fact that should all be ground plane surrounding that thing, anyway.  (You put ground plane / pour on your boards, right?...)

Yup, of course! Done too many boards to forget that thing now!

Keep everything away from the antenna end, especially signal traces and ground pour.  You usually want a keepout on the board in that region.  The datasheet probably talks about this.

Indeed it does, and I modified sparkfun's footprint to add a keepout on both the top and bottom, for parts, traces, and polyfills. I have 2mm clearance around the antenna, then surrounded with groundplane.

Cool, I'm not too concerned with HF noise, the rev A board I've been testing with has been working great, so if that ceramic resonator won't affect anything, it should be good to go!

Thanks for the tips folks!