> And as Tom pointed out, to be confident that your e-10 oscillator has > not drifted beyond e-10, you will need some means of periodically > comparing it to a better standard (which could be GPS or WWV, or a > better local standard such as a cesium or hydrogen maser > source). This will require some additional equipment (alternatively, > you could send the oscillator away for calibration periodically).
Here's an example. Assume someone wants a 1e-10 accurate frequency standard at home and they don't want to get too complicated, or don't want to become a time nut. A $100 rubidium oscillator [1] accurate to 1e-10 will drift in time by 8.6 microseconds. This can be *easily* measured against a $40 GPS/1PPS receiver [2] using a $1 microcontroller [3]. You get a hint of frequency error within an hour and solid data within a day. Just leave the GPS and MCU connected. You could log data once a day if you wanted to, even with pencil and paper. If the Rb has a frequency drift rate of 1e-11/month, you should expect to re-adjust the Rb once a year to maintain your +/-1e-10 spec. But that's all there is to it. Remember the birthday or wedding anniversary, change the batteries in your smoke detectors, and adjust your Rb. Note this is sort of a GPSDO; in this case the "disciplining" is done with a screwdriver once or twice a year. But it works for 99% of people looking for a frequency standard at home. It qualifies as low cost and low maintenance. It might be better to call it GPSTO, for a GPS Tracked Oscillator. In some cases knowing how much or little error your home frequency standard has today is all that's important. You don't actually need to adjust it every second, or minute like what a conventional GPSDO does under the hood. The short-term stability of a GPSTO is usually better than a GPSDO, anyway. /tvb [1] dozens to choose from on eBay [2] https://www.adafruit.com/products/746 [3] http://leapsecond.com/pic/ or do the equivalent with AVR, or Arduino, etc. _______________________________________________ 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.
