# What is the difference between Quartz Clock and Atomic Clock?

I want to know why an atomic clock is more accurate than a quartz one?

I understand that the atomic clock can give time to the nanosecond, say it is 10:30:24.123456789. If we have 1GHz quartz clock, then it means that it ticks 1 000 000 000 times every second, so it can give us an accuracy of 1/1000000000 = 0.000000001 second Am I wrong ?

• The difference is stability ([en.wikipedia.org/wiki/Allan_variance](Allen variance) is a keyword here). But I think this question is a bit off topic here.
– Ludo
Feb 15, 2022 at 10:55
• Also, I think 1 GHz quartz crystal doesn't exist; usually a smaller frequency quartz crystal output is frequency multiplied to reach 1 GHz. In that process errors may also get multiplied.
– AJN
Feb 15, 2022 at 12:03
• So, when I see my intel processor with a 2.2GHz clock, it means it has been multiplied to reach it. That means we don't have a 2.2GHz quartz clock in our microprocessor Feb 15, 2022 at 12:11
• and what about this part "so it can give us an accuracy of 1/1000000000 = 0.000000001 second Am I wrong ?" is it correct or not ? Feb 15, 2022 at 12:11
• Re Am I wrong ? Yes, you are wrong. The clock in your CPU more or less ticks at the advertised rate. The "more or less" is important for a number of reasons. Suppose you have two supposedly identical laptops. If you disconnect them from the internet so they cannot get Network Time Protocol (NTP) clock updates, you'll inevitably find that the clocks on the two laptops inevitably get out of sync. The "more or less" does not mean "exactly". The technical term is "within spec". I more or less tend to use "more or less" instead. Feb 15, 2022 at 14:31

Am I wrong ?

Yes, you are wrong. The clock in your CPU more or less ticks at the advertised rate. The "more or less" is important for a number of reasons. Suppose you have two supposedly identical laptops. If you disconnect them from the internet so they cannot get Network Time Protocol (NTP) clock updates, you'll inevitably find that the clocks on the two laptops inevitably get out of sync. The "more or less" does not mean "exactly". The technical term is "within spec". I more or less tend to use "more or less" instead.

All clocks, including atomic clocks, inevitably have a number of things that contribute to clock error. Some clocks run faster or slower than an ideal clock. (Note well: Humans have yet to create an ideal clock.) This is a clock's drift rate, which oftentimes varies with temperature. Crystal clocks tend to have a lot more drift than do atomic clocks. Even with a perfectly synchronized clocks, the time difference between one clock tick and the next on clock A versus clock B will vary. This is clock jitter. Crystal clocks have a lot more jitter than do atomic clocks.

Drift and jitter are only two of the many "features" (aka bugs) that infect real clocks.

Where this comes into play with regard to space exploration is that most computer systems used for flight software tend to have a very low jitter and very low drift clock that emits one pulse per second. This is so common that it has its own acronym, 1PPS.

• This reminds me of an infamous episode in the Shuttle Mission Simulator where, after the hardware contractor changed, the switch that synced all the different computers in the simulator to the same time got turned off. All the computers faulted down to internal time and slowly, slowly their clocks drifted apart. After a while, strange things started happening in the sim and got worse over time. Very difficult to figure out the problem. Feb 15, 2022 at 15:17
• @OrganicMarble Do have a link to more info on that? Feb 15, 2022 at 15:44
• @AlexHajnal spaceflight.training/black_friday.htm but depending on where your IP originates you may not be able to open it. Feb 15, 2022 at 16:16
• @OrganicMarble Thanks. (Loads fine for me.) Feb 16, 2022 at 5:20