Timekeeping on Mars or other worlds

[Andy44]

I was having a discussion with a friend at work who talks watches, and we were discussing how you would tell time on Mars.

Unlike, say, the Moon, which has a day nowhere near the same length as Earth's (about 29 days), Mars has a day very close to the same length as Earth: 24h 39m 35.244s. (NASA scientists have taken to calling Martian days "sols", which, to my thinking, may add more confusion, depending.)

So, it's reasonable to assume that an explorer spending several months on Mars will adapt a sleep cycle based on this slightly longer day.

It's also reasonable to assume that our explorer is not going to adopt a new base unit of time, but will continue using the SI second to measure time.

So here's the problem: The SI second is reverse-engineered from the Solar Day, which, before the atomic clock, was the standard unit of time for centuries. We divided the day into 24 hours, each of 60 minutes, each of 60 seconds. Then, when we realized that the Earth's rotation is not constant, we redefined the second as "the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium 133 atom".

Problem is, on Mars, the second does not divide evenly into a Martian solar day.

So I can build a watch that looks like an Earth watch, but runs slow enough so that a Martian noon will occur each sol at 12:00, but this means redefining hours, minutes, and seconds to Martian standards.

Or, I can stick with the SI second and live with a wierd ratio between seconds and Martian minutes or hours. But remember, the SI second does not divide evenly into a martian sol, so my watch may need often resetting and cannot be considered a precision clock.

What do you guys think? How would you measure time on Mars, while keeping up to speed with the SI system?

 

[SiberianTiger]

In fact, the time cycles on Earth are out of sync with the SI, too - a year is not a whole amount of seconds, a month is an arbitrary sequence of days about 30 days long in average. This causes inconvenience for any time span involving work, such as accounting - and that's why accountants devised a specific rules for financial calendars.

I believe that if people will ever live on Mars, they will have two calendars: one standard Earth, and another one Martian, following a complex computer generated pattern with leap days often used to introduce adjustments that keep it more or less in sync with Earth's weeks or months.

However, in case their facilities will be completely underground and they will rarely see the sunlight, then there is possibly no sense in having a special calendar and time keeping for the planet. One exiting should simply refer to the current data report, what time of day is outside before going, like we check weather forecasts.

 

[Kozak]

Sergey Lukyanenko in his novel Spectrum addresses this problem in such a way: one hour may consist of variable amount of time (61 minute 40 seconds in this case), while a day is still 24 hours. To round the values up - the last hour in the day may lack or have redundant seconds and fraction of seconds. So, 5 o'clock PM is time for tea regardless on which planet you are.

 

[SiberianTiger]

Sergey Lukyanenko in his novel Spectrum addresses this problem in such a way: one hour may consist of variable amount of time (61 minute 40 seconds in this case), while a day is still 24 hours. To round the values up - the last hour in the day may lack or have redundant seconds and fraction of seconds. So, 5 o'clock PM is time for tea regardless on which planet you are.

Out of the known planets, works for Mars only.

 

[Andy44]

What I'm thinking is how would you design a wristwatch that accurately tells you the local Mars time while still using the SI second. Reason being, you may use it as a stopwatch chronograph, and you're going to want SI seconds for that.

But you also want the watch set so that the sun rises the same time each day, which is pretty much the original function of a watch or a clock.

So what DarkWanderer says is interesting. We can redefine the Martian minute as having more than 60 seconds, or the hour as being more than 60 minutes, or better yet, redefine both so that they are devided into the same number of subunits, to minimize the number of dials on the watch dial.

Or...you just worry about minutes and hours, so that a Martian day is 24 martian hrs or 1440 martian minutes, and only count seconds on the stopwatch function.

---------- Post added at 04:37 PM ---------- Previous post was at 04:35 PM ----------

Oh, about other planets, this whole thing is pointless unless you are on a planet whose day is sufficiently close to 24 SI hours that people will set their sleep schedules around it.

Thus on the Moon you don't need to worry about this, because nobody will be awake or asleep for 2 weeks! They will just use GMT everywhere.

 

[TMac3000]

Why not just use GMT/UT regardless? Humans are on an Earth cycle: we go to bed when we get tired and get up when we are refreshed. Okay, so on Mars you go to bed when the sun is still shining. Pull the blinds

 

[Andy44]

Because if you work outside, say you are a geologist, you are going to want to do it when the sun is up.

---------- Post added at 05:00 PM ---------- Previous post was at 04:53 PM ----------

Okay, so applying a little math, if I divide a Martian sol into 24 equal units, each one is 3698.9685 SI sec long.

Taking the square root of that, I get a martian hour of 60.819 mars minutes, each of which is 60.819 SI sec.

Not very neat...will do some thinking and see what I come up with. I hate to resort to leap seconds, unless it's something you only need every few weeks or so.

 

[RisingFury]

Just measure time in seconds instead. A day lasts roughly 87000 seconds. Define an hour (Mour?) as 4000 seconds, which gives you 21.74 "Mours". Start counting from solar midnight and let the hour before the midnight have less seconds. Since everyone is sleeping at the time, it won't bother anyone.

 

[Hielor]

Why not just use GMT/UT regardless? Humans are on an Earth cycle: we go to bed when we get tired and get up when we are refreshed. Okay, so on Mars you go to bed when the sun is still shining. Pull the blinds

Humans living on Earth who are working on the Mars rovers already use the Sol system, which apparently wreaks havoc with their internal clock because their day is changing by ~30min relative to an Earth day each month.

As for the solution, I'm with RisingFury. You need to either have an odd amount of seconds every minute (which would just be...odd), an odd amount of minutes every hour (also odd), or one odd hour each day.

Since humans usually don't need to think about times longer than a day with that much precision (in their head, I mean), I'd propose something like having normal seconds/minutes and each day having 24 normal Earth hours and 1 extra "hour" which makes up for the extra time...

 

[2552]

What about having 25 hours, with the last hour being 37 minutes long?

Also, http://en.wikipedia.org/wiki/Darian_calendar

 

[Andy44]

Also, [URL="http://en.wikipedia.org/wiki/Darian_calendar"]http://en.wikipedia.org/wiki/Darian_calendar[/URL]

Very interesting, but doesn't get into the day itself.

The problem with just adding an extra 39m 35.244s "hour" to the day is that you have to represent this on a watch dial somehow, since that was the point of the original discussion.

A terrestrial watch dial divides into 12 neatly spaced increments, each of which divides into 5 neatly spaced increment for minutes and seconds. A martian watch with an extra 39m 35.244s is going to be kind of funny looking, don't you think?

 

[atuhalpa]

In Kim Stanley Robinson's Mars series, the colonists on Mars just engineered thier watches to stop for those 39 plus minutes and then start again.

 

[Linguofreak]

Very interesting, but doesn't get into the day itself.

The problem with just adding an extra 39m 35.244s "hour" to the day is that you have to represent this on a watch dial somehow, since that was the point of the original discussion.

A terrestrial watch dial divides into 12 neatly spaced increments, each of which divides into 5 neatly spaced increment for minutes and seconds. A martian watch with an extra 39m 35.244s is going to be kind of funny looking, don't you think?

The solution to the watch dial issue is simple:

Realize that analog watches are obsolete.

A digital watch can be programmed to deal with whatever oddities your timekeeping system may have. And "6" always means "6," rather than meaning "6" when refering to hours, and "30" when refering to minutes or seconds.

The second thing you need to do is have any interplanetary data or scheduling be based on just one time unit, which you then metricize and count from an epoch. Examples would be using seconds, kiloseconds, megaseconds, etc, and counting from the Unix epoch, or using Earth days, millidays, kilodays, etc, and counting by MJD or some similar system.

Planetary operations are then carried out according to a timekeeping system designed for the planet in question using non-standard units. All timekeeping devices are designed to display and convert between both your metricized standard time and your local time.

 

[Kozak]

Planetary operations are then carried out according to a timekeeping system designed for the planet in question using non-standard units. All timekeeping devices are designed to display and convert between both your metricized standard time and your local time.

I'd do with time of day expressed hexadecimal in seconds, but what with the other people?.. Timekeeping is not a challene for engineers, it's everyday need.

 

[Jarvitä]

I don't see the issue with using KSC/Baikonur time, as has always been the case. The length of the martian day is largely irrelevant, you can still arrange any activities that need to take place at day/night/second full moon after the winter solstice, regardless of the time system you use.

 

[Andy44]

The solution to the watch dial issue is simple:

Realize that analog watches are obsolete.

A digital watch can be programmed to deal with whatever oddities your timekeeping system may have. And "6" always means "6," rather than meaning "6" when refering to hours, and "30" when refering to minutes or seconds.

The second thing you need to do is have any interplanetary data or scheduling be based on just one time unit, which you then metricize and count from an epoch. Examples would be using seconds, kiloseconds, megaseconds, etc, and counting from the Unix epoch, or using Earth days, millidays, kilodays, etc, and counting by MJD or some similar system.

Planetary operations are then carried out according to a timekeeping system designed for the planet in question using non-standard units. All timekeeping devices are designed to display and convert between both your metricized standard time and your local time.

Ah, but mechanical watches are so much more elegant than a digital watch. And they don't need batteries and are less suceptible to radiation and charged particle damage, too. The old Omega from the Apollo days is still in use.

In fact, the clocks in hte Apollo vehicle cockpits were electronic, but not solid state. Rather, they were Accutron movements which used an electromagnetic tuning fork as a regulation device. Bulova stopped production of this type of movement in the mid-70s. But this is another type of movement which is hardier than a silicon chip.

Maybe I'm being unrealistic or stubborn, since I realize that you're right: a digital watch that just converts between two planets' time systems would be the simpler and cheaper route to go.

 

[tblaxland]

Ah, but mechanical watches are so much more elegant than a digital watch. And they don't need batteries and are less suceptible to radiation and charged particle damage, too. The old Omega from the Apollo days is still in use.

Not only are they more elegant, but they also give a better "feel" for the time relative to some other time in the past or future. For example, I know I have a meeting at 11am and I glance at my analogue watch and see that I have roughly one third of an hour until that meeting. For me this takes less "brain cycles" to process because I don't need to know the actual numbers, I just need to see the position of the hands relative to dial. Humans generally have a pretty good spatial perception (thanks to our evolution) and an analogue watch exploits that. I think this is the same reason why we simulate analogue gauges on digital instrument panels (eg, altitude tapes, ADI balls, etc).

 

[Andy44]

Yes, and why after some initial failed experiments, most car instruments are analog dials again.

 

[Linguofreak]

Yes, and why after some initial failed experiments, most car instruments are analog dials again.

But, strangely enough, most car clocks are digital...

 

[tblaxland]

But, strangely enough, most car clocks are digital...

Even in the car, I use my analogue watch in preference to the three other digital clocks I have at my disposal (car clock, mobile phone, PDA).

Getting back to the original topic, I would have 24 equal hours (3698.9685 s long) in the day each containing 60 equal minutes (61.6495 s long). I don't think that I can easily subjectively perceive the difference between a 60 s minute and 61.6495 s minute. As far as displaying that on a watch dial, I would have a main dial the same as usual but without a seconds hand. This effectively de-couples the length of the minutes from the number of seconds in them and you can glance at you watch and think "yep, I have 20 minutes until that meeting" and it will still "feel" the same as it did back on Terra. I would then have a seconds sub-dial that ticks in SI seconds for general counting purposes. You can add other sub-dials to your hearts content, eg, 24 hour UTC dial, stopwatch, sol number, alarm, etc.