4 ms·

Correct. Oscillators are subject to drift through a number of means, and they all have ridiculous effects too. https://news.ycombinator.com/item?id=37613523 htt

Correct. Oscillators are subject to drift through a number of means, and they all have ridiculous effects too. https://news.ycombinator.com/item?id=37613523 https://news.ycombinator.com/item?id=37613523

jrockway 3y ago
Temperature is the largest factor. Things like a DS3231[1] do really well compared to a basic non-compensated oscillator. I have been running some long-term experiments on a few that I have around and with some tuning got them to less than a second loss per year. But, they are super expensive compared to the basic ones (almost $5 each in quantity), so they aren't going to end up in your car where a 3 cent chip is possible to use instead. (I don't know what 5G / LTE chips cost these days, but if they're putting one in your car anyway, then they can probably get the time from that. But choose not to.) [1] https://www.analog.com/media/en/technical-documentation/data-sheets/DS3231.pdf https://www.analog.com/media/en/technical-documentation/data... Most interesting to me in all of my time experiments is looking at my clock frequency over time vs. the temperature. (NTP daemons aim to calculate your actual clock frequency; then they know how far off your internal time is from actual time.) You don't even need a temperature sensor, the clock rate is a perfect analogue.
- lazide 3y ago
  Voltage issues can also be a big problem, and cars have notoriously dirty electrical.
  - jrockway 3y ago
    Ahh, I bet that's true!
crote 3y ago
> A company i worked for wanted systems to have no more than 2ns of time drift between each other, in a network of +10 devices. At that point it's surprising they didn't just deploy a local "time network", with a single master clock distributing time via length-calibrated coax. Approaches like that are really common in television studios.
- CableNinja 3y ago
  It wasnt really the right environment for it, and they didnt even actually need that high of resolution, they couldve gotten away with 100ms drift and never noticed
SAI_Peregrinus 3y ago
Yeah, temperature is a big one. As is voltage supply stability. If you want really good short & medium term stability, it's hard (expensive) to do better than an Oven-Controlled Crystal Oscillator (OCXO). An OCXO has a crystal in an insulated chamber with a heater and a thermocouple. A control circuit uses the heater to keep the chamber at a consistent temperature. Over long periods these still drift. A cheaper alternative is a Temperature Compensated Crystal Oscillator (TCXO), that combines a control circuit with a crystal, a temperature sensor, and a ROM. The ROM contains a table of frequency errors that crystal had across temperature. The control circuit senses the temperature, reads the error value from the ROM, and tries to correct the crystal's oscillation frequency to compensate for that amount of error. Less accurate and less stable than an OCXO, but much smaller, much lower power, and much cheaper. For longer-term stability you want an atomic clock, probably a Cesium clock or Hydrogen Maser. GNSS (GPS, Galileo, GLONASS, Baidu, etc) satellites have atomic clocks on board, and broadcast that time. GNSS is based around very precisely measuring the differences in received times from various clocks to calculate location, so it can also be used as a way to get a local time reference with good long-term stability. Unfortunately GNSS does have rather poor short-term stability compared to an OCXO, so GNSS alone isn't perfect. But it's common and reasonably inexpensive to create a "GPS-Disciplined OCXO" where a GPS unit corrects for the long-term drift of an OCXO, but the OCXO provides the actual output signal (thus gaining the good short & medium-term stability). The NIST SP 1065 Handbook of Frequency Stability Analysis[1] is a go-to text on measuring clock sources. [1] https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication1065.pdf https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpubli...

jrockway 3y ago

Temperature is the largest factor. Things like a DS3231[1] do really well compared to a basic non-compensated oscillator. I have been running some long-term experiments on a few that I have around and with some tuning got them to less than a second loss per year. But, they are super expensive compared to the basic ones (almost $5 each in quantity), so they aren't going to end up in your car where a 3 cent chip is possible to use instead. (I don't know what 5G / LTE chips cost these days, but if they're putting one in your car anyway, then they can probably get the time from that. But choose not to.) [1] https://www.analog.com/media/en/technical-documentation/data-sheets/DS3231.pdf https://www.analog.com/media/en/technical-documentation/data... Most interesting to me in all of my time experiments is looking at my clock frequency over time vs. the temperature. (NTP daemons aim to calculate your actual clock frequency; then they know how far off your internal time is from actual time.) You don't even need a temperature sensor, the clock rate is a perfect analogue.

lazide 3y ago
Voltage issues can also be a big problem, and cars have notoriously dirty electrical.
- jrockway 3y ago
  Ahh, I bet that's true!

crote 3y ago

> A company i worked for wanted systems to have no more than 2ns of time drift between each other, in a network of +10 devices. At that point it's surprising they didn't just deploy a local "time network", with a single master clock distributing time via length-calibrated coax. Approaches like that are really common in television studios.

CableNinja 3y ago
It wasnt really the right environment for it, and they didnt even actually need that high of resolution, they couldve gotten away with 100ms drift and never noticed

SAI_Peregrinus 3y ago

Yeah, temperature is a big one. As is voltage supply stability. If you want really good short & medium term stability, it's hard (expensive) to do better than an Oven-Controlled Crystal Oscillator (OCXO). An OCXO has a crystal in an insulated chamber with a heater and a thermocouple. A control circuit uses the heater to keep the chamber at a consistent temperature. Over long periods these still drift. A cheaper alternative is a Temperature Compensated Crystal Oscillator (TCXO), that combines a control circuit with a crystal, a temperature sensor, and a ROM. The ROM contains a table of frequency errors that crystal had across temperature. The control circuit senses the temperature, reads the error value from the ROM, and tries to correct the crystal's oscillation frequency to compensate for that amount of error. Less accurate and less stable than an OCXO, but much smaller, much lower power, and much cheaper. For longer-term stability you want an atomic clock, probably a Cesium clock or Hydrogen Maser. GNSS (GPS, Galileo, GLONASS, Baidu, etc) satellites have atomic clocks on board, and broadcast that time. GNSS is based around very precisely measuring the differences in received times from various clocks to calculate location, so it can also be used as a way to get a local time reference with good long-term stability. Unfortunately GNSS does have rather poor short-term stability compared to an OCXO, so GNSS alone isn't perfect. But it's common and reasonably inexpensive to create a "GPS-Disciplined OCXO" where a GPS unit corrects for the long-term drift of an OCXO, but the OCXO provides the actual output signal (thus gaining the good short & medium-term stability). The NIST SP 1065 Handbook of Frequency Stability Analysis[1] is a go-to text on measuring clock sources. [1] https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication1065.pdf https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpubli...