); SAEximRunCond expanded to false Errors-To: [EMAIL PROTECTED] RETRY There was an interesting homebrew proton-precession magnetometer in Circuit Cellar magazine a few months back. Not very exotic compared to Cs and Rb tech, but somewhat more accessible to experimenters. You fill two 16-ounce plastic bottles with either water or kerosene, wind some wire around them both, and connect the coils in series at the bottom to cancel external AC fields. Then you magnetize the coils for a short time depending on what liquid you used. When you cut the current off, the coil will ring down at a frequency proportional to the local magnetic field. Accuracy and resolution (looking back at the article) both appear to be in the sub-nT range.
ObTimeNuts content: apparently most PPMs use reciprocal counters to measure the ringing frequency, but the author of this article used a trick I wasn't familiar with. He oversamples the signal and divides the amplitude difference observed between successive pairs of even-numbered sample points by that observed at successive odd-numbered points. Then, the atan() of that ratio yields an absolute phase. Given a fixed sampling clock, the change in phase from one 4-sample set to the next yields the frequency of the waveform, removing any DC offset in the bargain. Since each cycle is measured with 4 points instead of 2 points with a traditional zero-crossing detector, you apparently get better noise immunity as well (I'd guess 6 dB worth.) It reminded me of how the TSC-5120A data sheet says they measure phase with an arctan function. I don't immediately see why this is better than an FFT in either the TSC-5120A's case or the magnetometer's... except that it's more amenable to implementation on a low-powered microcontroller in the Circuit Cellar project. -- john, KE5FX > NIST is making it small, which is new, but the technology is > well-established. See, for example: > > http://www.geometrics.com/magnetometers/magnetometers.html > _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.