QUOTE (Matt Lenda @ Aug 6 2011, 04:40 PM)
Yes and no. That's an extraordinarily complex issue, not least because we have to ask the question, "What clock?" and the question, "Relative to what?"
We have several ways of mitigating the drift, none of which are trivial. There's no real *reset* button to do it. (And for good reason.)
We have seen less drift of this particular clock when the rover is warmer. The theory is well supported by the data, but we have yet to implement the fix in tactical activities. It's low on our radar.
I'll stop talking before I get in over my head.
-m
Interesting stuff Matt! I've had to deal with oscillators in my line of work. Most issues have been around the Real Time Clock. I remember one customer routed their digital 32K output right underneath the 32K crystal, which is a big no-no. Caused so much feedback noise into the crystal that it was running at around 130 kHz. Another funny issue was one customer complaining about our RTC being off 2 minutes after "only" a month. We had to explain to them that the 50ppm for a typical 32kHz crystal meant that over a month, you could be up to 2 minutes off! We did have options for calibrating the crystal, but they didn't do that during their manufacturing run because it took too long. LOL, fun times.
I've also had to do some wireless stuff as well, WiFi and GSM. For the non-engineers, frequency crystals can be pretty accurate (As long as your design isn't noisy), typically 50 to 200ppm. Their biggest variance short-term is due to temperature, as Opportunity is experiencing, or due to aging. For items requiring precision, what is typically done is using a tiny chip to adjust the frequency based on the ambient temperature, reducing the frequency variability significantly, usually to less than 1ppm. These are called Temperature Compensated Crystal Oscillators, or TCXOs. For the ultimate in frequency control, the crystal is stuck inside an oven, guaranteeing there is no temperature change at all, and getting < .001ppm, called Oven-Controlled Crystal Oscillators, or OCXO. I think the "ultra stable oscillators" discussed for radio science uses these types of oscillators to provide the stable frequency for measuring changes due to gravity.
For Oppy it sounds like the oven itself might be aging and not keeping the crystal at the ideal temperature. Another possibility might be a decoupling capacitor aging, or solder itself starting to exhibit increased resistance due to temp. cycles.
For the curious, there is some more information here
http://www2.electronicproducts.com/Underst...b2005-html.aspx