Hello, I just got my NerdKit and got through the first stage of connecting the screen. When I plugged in the battery, everything worked fine. But I happened to be listening to the radio at the time (88.5 FM), and the interference was fierce! Can anyone explain it? I assume something is radiating at the crystal frequency. What would you do to stop it? Thanks, Rob

Rob,

That's an interesting situation. I can't explain what's going on exactly (I'm not as up on the physics as I'd like to be), but I can tell you that the NerdKit clock frequency of 14.7456, multiplied by 6, is 88.4736, very close to where you were tuned. This supports your theory that it is linked to the crystal oscillator. The oscillator is probably being noisy. Does your NerdKit work properly and communicate over serial? (Can you load new programs?). My advice is to email NerdKits, they'll probably look at replacing it with a new crystal. In the mean time, try the standard techniques of repositioning the radio, the antenna, the NerdKit, etc. Make sure that everything that plugs in is properly grounded. Just to make sure, are you in the US? (Different voltages, hertz, etc). Let us know how it goes. Good luck, John

Proximity is likely a lot of the culprit. The kits are largely un-shielded and as a result, emit a cacophony of signals in the RF spectrum. There are acceptable levels and limitations imposed by the FCC and I think it is probably within the limits. If the interference decreases with distance, it would do so exponentially and the range would be limited.

Hi mongo & John, Thanks for the replies. I haven't gotten to the point of loading programs yet, but everything seems to be working ok. One thing that occurred to me: looking at the pictures in the manual, they seem to have cut the leads to the crystal so it rests right on the board. I have probably about 1/4" showing on mine, so they may be the culprits, acting like little transmitting antennas. I might cut them shorter, or I think putting ferrite beads on them can help too(?) The factor of 6 that John points out is interesting. That means the crystal wavelength is 6 times longer than the radio wave's at 88.5 MHz. If I recall my physics correctly, an even multiple like that could get through the tuning circuitry. (Is that right?) I wonder if the length of my leads is also some multiple of the wavelengths? Thanks again, Rob

I'm not going to respond, simply because I don't understand the physics well enough. Did moving the radio help?

no haven't tried moving it yet, but will when I get home. Thanks again.

Every wire and conductor can act as an antenna and bleed RF energy all over the place. FCC part 95 addresses these issues but it has been a long time since I have read it. The best thing to do is keep the leads as short as you can and relocate the radio or the kit.

Hi Rob,

I think johnh is definitely on the right track with the 14.7456*6 = 88.47 MHz calculation!

An FM radio receiver is designed to "lock on" to a frequency that is roughly within about +/- 100kHz of the nominal 88.5 MHz setpoint. And when I say "lock on", I mean it -- it uses a phase locked loop to try to take a super weak signal and follow it. It does not take much of a signal being generated a few feet from a radio to get it to lock on.

The crystal's nominal frequency is 14.7456 MHz, but the "input" to the crystal (being generated by the microcontroller) is more like a square wave than a sine wave. If it's a square wave means that it has frequency components at lots of multiples of 14.7456 MHz. For an perfect square wave, that's actually only for odd multiples (1,3,5,7...) of the base frequency, which would skip 6. However, if there's some asymmetric distortion going on (maybe different transistor drive strengths inside the microcontroller), you can get the even multiples (2,4,6,8...) as well.

The first thing I'd check would be to try programming mode (which basically is just idle unless you're actively uploading) and see if that quiets things down! You may find that the connections to components like the LCD, rather than just the CPU running, may be the cause of your issues.

Trimming the crystal leads is a good idea, although with the solderless breadboard, there's still big spring contacts and lots of area there. Add the 22pF capacitors (included with kit) between each crystal pin and ground. If all else fails, you can try a metal enclosure. You may get much of the metal enclosure effect simply by placing the breadboard on top of a metal plate/pan.

In any case, remember that radios are designed to lock on to even super-weak signals in the tuned range. Reducing the level of emissions may be possible, but short of using a sealed metal enclosure, it may be hard to make it go away completely. I haven't been able to repeat this issue yet, but if you give it a try, let us know how each of these changes affects the radio signal!

Mike

I agree. PLL's are sensitive and if you hit the right harmonic freq, it can pick up unwanted signals and lock on to them. Lots of options to pick and choose from.

Hi Mike & mongo,

Thanks for the suggestions. I'm a little lost in the details, but think I get the drift of it. I'll try to get started on the temperature sensor project in the next few days, and let you know how it goes. I'll try the capacitor and metal plate too, just out of curiosity.

Thanks again, Rob

I've been playing with the temperature sensor for a while now and kind of forgot about this problem (I'm also listening to satellite radio more, which isn't bothered by it at all). Anyway, there doesn't seem to be any very noticeable interference now. So I'm wondering if it didn't have something to do with the line of scrolling text in the pre-loaded program, which I don't have anymore.