Simon's Worldmap - A question on sunset countdown time
Today I displayed your Worldmap on a second computer during the ARRL contest and I observed that the minutes above the Sunset are one minute ahead of the remaining time. E.g. The UTC time is 11:59 and the sunset time is 16:18
When the UTC time switches to 12:00 the remaining time displayed above the sunset time should display 4h and 18 minutes. However it displays 4h and 17 minutes instead. Is this made on purpose ?
Again many thanks for the wonderful work and the many hours spent on this project. I've contributed to some food for your dog.
Ezio - F5MWA / K3EK
Sent: 10 December 2022 16:06
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
Today I displayed your Worldmap on a second computer during the ARRL contest and I observed that the minutes above the Sunset are one minute ahead of the remaining time. E.g. The UTC time is 11:59 and the sunset time is 16:18
When the UTC time switches to 12:00 the remaining time displayed above the sunset time should display 4h and 18 minutes. However it displays 4h and 17 minutes instead. Is this made on purpose ?
Again many thanks for the wonderful work and the many hours spent on this project. I've contributed to some food for your dog.
Ezio - F5MWA / K3EK
--
You answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences
you have would be a good test. If you use the multiple windows
feature alt-win-arrow moves from one main window to the other so
both can be full screen.
{o.o}
Today I displayed your Worldmap on a second computer du...
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
You answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
Today I displayed your Worldmap on a second computer du...
--
That's exactly what I run too.
W1HHO
-- All the Best from the woods of Maine.If you run Meinberg NTP you can be very close to atomic time.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeYou answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
On 20221210 08:06:54, f5mwa wrote:
Today I displayed your Worldmap on a second computer du...
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
That is "if". Out of the box windows is loosey goosey updating time once a day and drifting through the next 24 hours. So that's why I suggested the single computer check. That ALSO must deal with how often the programs reach out to update time. If they reach out every second, as with a timer set to 1 second, you can expect the two machines to read different values a goodly percentage of the time depending on when each started their timer requests.
Start #1 process at 0.01 after the computer's 1 second tick and start #2 process at 0.85 after the 1 second tick and most of the time they will read 1 second off from each other.
{o.o}
If you run Meinberg NTP you can be very close to atomic time.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeYou answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
On 20221210 08:06:54, f5mwa wrote:
Today I displayed your Worldmap on a second computer du...
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
And I just showed you an example that can lead to 1 second plus
or minus on the readings between process time updates on the
screen. If you demand better real accuracy on screen then perhaps
10 ms ticks, read time, if seconds ticks over update screen. Then
you'd differ by a second for up to 10 or 20 ms if you are lucky.
Process priorities and task switching times can add more update
time uncertainty,
I have found simple ways to get a signal in a process once every n ms. I've not found a simple low "CPU cycle cost" way to get signal when the seconds digit ticks over. (Millisecond timer flips from decimal XXXXX999 to XXXXY000 with Y = X+1.)
{^_^}
That's exactly what I run too.
W1HHO
On 12/10/2022 13:16, Simon Brown wrote:
-- All the Best from the woods of Maine.If you run Meinberg NTP you can be very close to atomic time.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeYou answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
On 20221210 08:06:54, f5mwa wrote:
Today I displayed your Worldmap on a second computer du...
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
Sent: 10 December 2022 21:33
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
And I just showed you an example that can lead to 1 second plus or minus on the readings between process time updates on the screen. If you demand better real accuracy on screen then perhaps 10 ms ticks, read time, if seconds ticks over update screen. Then
you'd differ by a second for up to 10 or 20 ms if you are lucky. Process priorities and task switching times can add more update time uncertainty,
I have found simple ways to get a signal in a process once every n ms. I've not found a simple low "CPU cycle cost" way to get signal when the seconds digit ticks over. (Millisecond timer flips from decimal XXXXX999 to XXXXY000 with Y = X+1.)
{^_^}
That's exactly what I run too.
W1HHO
On 12/10/2022 13:16, Simon Brown wrote:
-- All the Best from the woods of Maine.If you run Meinberg NTP you can be very close to atomic time.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeYou answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
On 20221210 08:06:54, f5mwa wrote:
Today I displayed your Worldmap on a second computer du...
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
--
Hi Joanne,
Have you used Meinberg? As I understand it, it continually (not sure of the interval, but very often) monitors the time signal from multiple NTP servers and also tracks the deviation of the PC time from the server time. It track that deviation, and ‘continually’ calculates the deviation and applies a correction to the system time to make it as close as possible to NIST/standard time. It also stores the statistics for the deviation and correction factor. Based on what I typically see, the deviation from UTC is usually within ±0.002 sec (± 2 msec), and the clock frequency correction is generally around -4.5 ± 0.2 ppm.
Bottom line… I feel my system time is relatively accurate to with a few milliseconds. If you ran Meinberg on each PC, it should track each deviation correction and clock frequency deviation for each individual computer, and adjust each computer accordingly to keep each one within a few milliseconds of true UTC. As such, I’d guess two PCs should be within about 10 msec of one another.
Granted, if you restart/reboot either PC, the system time will likely deviate ‘significantly’ (perhaps up to a few seconds) from UTC until Meinberg does its magic and resyncs the PC to UTC.
Thanks for your insights.
Ken – WBØOCV
Sent: Saturday, December 10, 2022 04:27 PM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
That is "if". Out of the box windows is loosey goosey updating time once a day and drifting through the next 24 hours. So that's why I suggested the single computer check. That ALSO must deal with how often the programs reach out to update time. If they reach out every second, as with a timer set to 1 second, you can expect the two machines to read different values a goodly percentage of the time depending on when each started their timer requests.
Start #1 process at 0.01 after the computer's 1 second tick and start #2 process at 0.85 after the 1 second tick and most of the time they will read 1 second off from each other.
{o.o}
Link to the multimedia stuff, timeapi.h for the time related api.
It allows setting system interrupt time to 1ms (or more.) But,
then, you know about timeGetTime() and its cousins for 1ms timing
and so forth. The precision timers historically are very CPU
intensive for reading their time. But they do give 100 ns timing
resolution, perhaps from a much lower precision 3.579... MHz cheap
NTSC color crystal. I'd expect better; but, would not bet on it.
These timers are often used in things like DirectShow or its newer
and even more arcane equivalent.
Linux is somewhat better for setting up precision timers. But, that requires compiling it with the right options, not often used in official distros.
{^_^}
It is possible. I run a 100ms timer, when the second changes I update the display. There are low-level calls you can use to get the tick count, these are often used in games.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 21:33
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeAnd I just showed you an example that can lead to 1 second plus or minus on the readings between process time updates on the screen. If you demand better real accuracy on screen then perhaps 10 ms ticks, read time, if seconds ticks over update screen. Then you'd differ by a second for up to 10 or 20 ms if you are lucky. Process priorities and task switching times can add more update time uncertainty,
I have found simple ways to get a signal in a process once every n ms. I've not found a simple low "CPU cycle cost" way to get signal when the seconds digit ticks over. (Millisecond timer flips from decimal XXXXX999 to XXXXY000 with Y = X+1.)
{^_^}
On 20221210 10:43:16, Calder Latham wrote:
That's exactly what I run too.
W1HHO
On 12/10/2022 13:16, Simon Brown wrote:
-- All the Best from the woods of Maine.If you run Meinberg NTP you can be very close to atomic time.
From: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io> on behalf of jdow via groups.io <jdow@...>
Sent: 10 December 2022 18:08
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown timeYou answered your own question. There is no guarantee that two computers are closer than a couple seconds to each other unless special measures are taken. And that does not account for machine activity pushing off updating clocks for higher priority processes.
Running both on the same computer to see what timing differences you have would be a good test. If you use the multiple windows feature alt-win-arrow moves from one main window to the other so both can be full screen.
{o.o}
On 20221210 08:06:54, f5mwa wrote:
Today I displayed your Worldmap on a second computer du...
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
--
- + - + -Please use https://forum.sdr-radio.com:4499/ when posting questions or problems.
It is a compile of "ntp" for Windows. I've been using ntp on Linux since 1997 or so. On the PC there is an old implementation of ntp that I've been using since I learned of it, WAY before Symmetricom's implementation or Meinberg's implementation. So, the answer is "no; but, I am well aware of it and using something roughly equivalent."
Log entries:
2022/12/10 23:07:32 UTC [8264/8336]: Requesting time from server:
XXXz
2022/12/10 23:07:32 UTC [8264/8336]: clock adjustment: 0.001
seconds
2022/12/10 23:07:32 UTC [8264/8336]: round trip delay: -0.000
seconds
2022/12/10 23:07:32 UTC [8264/8336]: server stratum: 2
2022/12/10 23:08:32 UTC [8264/8336]: Requesting time from server:
XXX
2022/12/10 23:08:32 UTC [8264/8336]: clock adjustment: 0.000
seconds
2022/12/10 23:08:32 UTC [8264/8336]: round trip delay: 0.001
seconds
2022/12/10 23:08:32 UTC [8264/8336]: server stratum: 2
2022/12/10 23:09:32 UTC [8264/8336]: Requesting time from server:
XXX
2022/12/10 23:09:32 UTC [8264/8336]: clock adjustment: 0.000
seconds
2022/12/10 23:09:32 UTC [8264/8336]: round trip delay: 0.001
seconds
2022/12/10 23:09:32 UTC [8264/8336]: server stratum: 2
{^_-}
Hi Joanne,
Have you used Meinberg? As I understand it, it continually (not sure of the interval, but very often) monitors the time signal from multiple NTP servers and also tracks the deviation of the PC time from the server time. It track that deviation, and ‘continually’ calculates the deviation and applies a correction to the system time to make it as close as possible to NIST/standard time. It also stores the statistics for the deviation and correction factor. Based on what I typically see, the deviation from UTC is usually within ±0.002 sec (± 2 msec), and the clock frequency correction is generally around -4.5 ± 0.2 ppm.
Bottom line… I feel my system time is relatively accurate to with a few milliseconds. If you ran Meinberg on each PC, it should track each deviation correction and clock frequency deviation for each individual computer, and adjust each computer accordingly to keep each one within a few milliseconds of true UTC. As such, I’d guess two PCs should be within about 10 msec of one another.
Granted, if you restart/reboot either PC, the system time will likely deviate ‘significantly’ (perhaps up to a few seconds) from UTC until Meinberg does its magic and resyncs the PC to UTC.
Thanks for your insights.
Ken – WBØOCV
From: jdow
Sent: Saturday, December 10, 2022 04:27 PM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
That is "if". Out of the box windows is loosey goosey updating time once a day and drifting through the next 24 hours. So that's why I suggested the single computer check. That ALSO must deal with how often the programs reach out to update time. If they reach out every second, as with a timer set to 1 second, you can expect the two machines to read different values a goodly percentage of the time depending on when each started their timer requests.
Start #1 process at 0.01 after the computer's 1 second tick and start #2 process at 0.85 after the 1 second tick and most of the time they will read 1 second off from each other.
{o.o}
There is a misunderstanding here. I did not want to compare the time displayed on two Worldmaps running on two different PCs. I only had mentioned my Contest setup. On one computer I had N1MM with SRC and CWSkimmer server and on the other I had the Simon's Worldmap. Now, when you look at the time since sunset the minutes displayed are ok whereas for the time before sunset the minutes are behind of 1 minute. Am I wrong ? Has it something to do with the routine you use for the countdown ?
Here the screenshot of this morning
Thanks again for the nice piece of software, a sort of Christmas present I believe.......
73 to all
Ezio - F5MWA / K3EK
Sent: 12 December 2022 08:01
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
There is a misunderstanding here. I did not want to compare the time displayed on two Worldmaps running on two different PCs. I only had mentioned my Contest setup. On one computer I had N1MM with SRC and CWSkimmer server and on the other I had the Simon's Worldmap. Now, when you look at the time since sunset the minutes displayed are ok whereas for the time before sunset the minutes are behind of 1 minute. Am I wrong ? Has it something to do with the routine you use for the countdown ?
Here the screenshot of this morning
Thanks again for the nice piece of software, a sort of Christmas present I believe.......
73 to all
Ezio - F5MWA / K3EK
--
maybe cause the actual minute is already running ;-)
dg9bfc sigi
Hello Simon (and Jdow)
There is a misunderstanding here. I did not want to compare the time displayed on two Worldmaps running on two different PCs. I only had mentioned my Contest setup. On one computer I had N1MM with SRC and CWSkimmer server and on the other I had the Simon's Worldmap. Now, when you look at the time since sunset the minutes displayed are ok whereas for the time before sunset the minutes are behind of 1 minute. Am I wrong ? Has it something to do with the routine you use for the countdown ?
Here the screenshot of this morning
Thanks again for the nice piece of software, a sort of Christmas present I believe.......
73 to all
Ezio - F5MWA / K3EK
Hello all,
Simon is right on the mark. Different software programs can use different equations for calculating the time of sunrise and sunset. If the equations are scientifically sound and properly vetted, there can still be slight differences between the results they produce, certainly enough to result in differences of tens of seconds, or perhaps even up to and over 1 minute. I don’t think any of the equations are necessarily “wrong”, but some might be more “correct” than others. On top of that, the equations usually apply to a given astronomical epoch, so can produce slightly different results if you’re on the cusp of an epoch. In addition, some equations may be based on slightly different assumptions regarding polar and equatorial diameters of the Earth, and different assumptions regarding the curvature of the Earth surface. Then you can factor in assumptions of the precise orbital position of the Earth around the Sun for a given date. All of these factors can affect the precise estimation of sunrise/sunset. Again, the results may be very precise relative to the equation used, but might not agree exactly with another program that uses different equations which use slightly different constants and factors.
For example, here in the USA one can go to the NOAA (National Oceanographic and Atmospheric Administration) website and drill down to find data for sunrise/sunset for a specific location. I would trust that the equations used by NOAA are ‘correct’. Same with some of the propagation websites. There is useful program/app called LunaSolCal that can also calculate sun and moon data for a specific location, which I find to be very good and useful. If one checks a weather forecast program (Weather.com, Accuweather.com, WeatherBug, etc.), these will also display sunrise/sunset for a given location, and are also likely using a vetted source for their data. And I have no doubt Simon used a commonly approved/vetted set of equations to calculate solar and lunar data, which may or may not be the same as those used by other “authorities”. It may produce slightly different results, but technically I don’t view the results as “wrong” (they are indeed correct based on the mathematics of the equation used).
Also another thing to look out for. Many of the web-based sources of sunrise/sunset data (e.g. local weather forecasts) will display the data for a closest city in their database, as opposed to your actual QTH. That may be tens of miles distant east/west or north/south from the actual QTH where the user is located. That 10 or 15 or 20 miles can easily account for a difference in ±1 minute. Since Simon’s World Map uses a precise QTH based on whatever you provide in the setup, it should produce accurate sunrise/sunset times that are correct for the mathematics inherent in the equations Simon chose to use. Will Simon’s values agree exactly with a NOAA prediction for the same latitude/longitude? Maybe, maybe not, but I suspect the results would agree within ±1 minute.
Everyone have a great day.
Ken
Sent: Monday, December 12, 2022 03:18 AM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
It depends how the other software calculates sunrise / sunset. I round to the nearest minute or 10 seconds. I believe my calculation to be correct.
As an addendum to my earlier post –
Some predictors of sunrise/sunset also employ a atmospheric diffraction correction to adjust the predicted time of sunrise/sunset for the bending of light through the atmosphere. That can also result in differences between calculated and perceived time of sunrise/sunset, and also for differences between predictions that do or do not apply such a correction.
Here is a link to the NOAA website that describes some of the information related to calculating solar data:
https://gml.noaa.gov/grad/solcalc/calcdetails.html
I’d encourage people to consider downloading the NOAA_Solar_Calculations_Year.xls file. This spreadsheet provides all the information one could want regarding time of sunrise, sunset, solar noon, declination, elevation, etc. for any given day of the year for a specific geographic location. Input the latitude and longitude of your QTH, and you should come up with the necessary information for your location. I especially find the time of solar noon to be very useful. At the time of solar noon, the sun is due-South (180.0°) of your location (opposite in the Southern hemisphere), and will be casting a shadow toward due-North (0.0°). This can be invaluable in placing a sighting post that is due-North of your tower, weather station, etc. to allow for precise directional alignment of an antenna rotor, weather sensor, etc. Although using a compass with proper declination correction will get you close enough, this is a preferred method for alignment sighting used by many agencies.
As an aside, I had downloaded an older (circa 1990s) pedigreed NOAA calculation methodology for solar calculations. That model was vetted and approved by NOAA, EPA, NWS, DOD, etc. at the time. I have used it regularly since then. Today I downloaded the “latest” NOAA spreadsheet, and ran both spreadsheets for my exact latitude and longitude. Not surprisingly, the two models agreed with one another within about ±10 seconds, although there were a couple of sunset times that deviated from one another by about 30-seconds, as highlighted below. Here’s a rundown for the two models for the equinoxes and solstices for my QTH for this year, with all times referenced to UTC:
Date | Sunrise | Sunset | Noon | |||
Old | New | Old | New | Old | New | |
3/20/2022 | 10:45:32 | 10:45:26 | 22:54:27 | 22:53:57 | 16:49:38 | 16:49:41 |
6/21/2022 | 09:08:11 | 9:08:13 | 00:20:00 | 0:19:53 | 16:44:06 | 16:44:03 |
9/20/2022 | 10:27:40 | 10:27:44 | 22:42:54 | 22:43:40 | 16:35:38 | 16:35:42 |
12/21/2022 | 12:05:38 | 12:05:38 | 21:15:12 | 21:15:00 | 16:40:25 | 16:40:19 |
For what it’s worth, Simon’s World Map provides sunrise/sunset times that closely align with the NOAA models, at least for today’s date.
I hope everyone has a great day.
Ken
Sent: Monday, December 12, 2022 03:18 AM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
It depends how the other software calculates sunrise / sunset. I round to the nearest minute or 10 seconds. I believe my calculation to be correct.
Sent: 12 December 2022 14:54
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
As an addendum to my earlier post –
Some predictors of sunrise/sunset also employ a atmospheric diffraction correction to adjust the predicted time of sunrise/sunset for the bending of light through the atmosphere. That can also result in differences between calculated and perceived time of sunrise/sunset, and also for differences between predictions that do or do not apply such a correction.
Here is a link to the NOAA website that describes some of the information related to calculating solar data:
https://gml.noaa.gov/grad/solcalc/calcdetails.html
I’d encourage people to consider downloading the NOAA_Solar_Calculations_Year.xls file. This spreadsheet provides all the information one could want regarding time of sunrise, sunset, solar noon, declination, elevation, etc. for any given day of the year for a specific geographic location. Input the latitude and longitude of your QTH, and you should come up with the necessary information for your location. I especially find the time of solar noon to be very useful. At the time of solar noon, the sun is due-South (180.0°) of your location (opposite in the Southern hemisphere), and will be casting a shadow toward due-North (0.0°). This can be invaluable in placing a sighting post that is due-North of your tower, weather station, etc. to allow for precise directional alignment of an antenna rotor, weather sensor, etc. Although using a compass with proper declination correction will get you close enough, this is a preferred method for alignment sighting used by many agencies.
As an aside, I had downloaded an older (circa 1990s) pedigreed NOAA calculation methodology for solar calculations. That model was vetted and approved by NOAA, EPA, NWS, DOD, etc. at the time. I have used it regularly since then. Today I downloaded the “latest” NOAA spreadsheet, and ran both spreadsheets for my exact latitude and longitude. Not surprisingly, the two models agreed with one another within about ±10 seconds, although there were a couple of sunset times that deviated from one another by about 30-seconds, as highlighted below. Here’s a rundown for the two models for the equinoxes and solstices for my QTH for this year, with all times referenced to UTC:
|
Date |
Sunrise |
Sunset |
Noon |
|||
|
Old |
New |
Old |
New |
Old |
New |
|
|
3/20/2022 |
10:45:32 |
10:45:26 |
22:54:27 |
22:53:57 |
16:49:38 |
16:49:41 |
|
6/21/2022 |
09:08:11 |
9:08:13 |
00:20:00 |
0:19:53 |
16:44:06 |
16:44:03 |
|
9/20/2022 |
10:27:40 |
10:27:44 |
22:42:54 |
22:43:40 |
16:35:38 |
16:35:42 |
|
12/21/2022 |
12:05:38 |
12:05:38 |
21:15:12 |
21:15:00 |
16:40:25 |
16:40:19 |
For what it’s worth, Simon’s World Map provides sunrise/sunset times that closely align with the NOAA models, at least for today’s date.
I hope everyone has a great day.
Ken
From: Simon Brown
Sent: Monday, December 12, 2022 03:18 AM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
It depends how the other software calculates sunrise / sunset. I round to the nearest minute or 10 seconds. I believe my calculation to be correct.
--
Sent: 12 December 2022 14:54
To: main@SDR-Radio.groups.io <main@SDR-Radio.groups.io>
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
As an addendum to my earlier post –
Some predictors of sunrise/sunset also employ a atmospheric diffraction correction to adjust the predicted time of sunrise/sunset for the bending of light through the atmosphere. That can also result in differences between calculated and perceived time of sunrise/sunset, and also for differences between predictions that do or do not apply such a correction.
Here is a link to the NOAA website that describes some of the information related to calculating solar data:
https://gml.noaa.gov/grad/solcalc/calcdetails.html
I’d encourage people to consider downloading the NOAA_Solar_Calculations_Year.xls file. This spreadsheet provides all the information one could want regarding time of sunrise, sunset, solar noon, declination, elevation, etc. for any given day of the year for a specific geographic location. Input the latitude and longitude of your QTH, and you should come up with the necessary information for your location. I especially find the time of solar noon to be very useful. At the time of solar noon, the sun is due-South (180.0°) of your location (opposite in the Southern hemisphere), and will be casting a shadow toward due-North (0.0°). This can be invaluable in placing a sighting post that is due-North of your tower, weather station, etc. to allow for precise directional alignment of an antenna rotor, weather sensor, etc. Although using a compass with proper declination correction will get you close enough, this is a preferred method for alignment sighting used by many agencies.
As an aside, I had downloaded an older (circa 1990s) pedigreed NOAA calculation methodology for solar calculations. That model was vetted and approved by NOAA, EPA, NWS, DOD, etc. at the time. I have used it regularly since then. Today I downloaded the “latest” NOAA spreadsheet, and ran both spreadsheets for my exact latitude and longitude. Not surprisingly, the two models agreed with one another within about ±10 seconds, although there were a couple of sunset times that deviated from one another by about 30-seconds, as highlighted below. Here’s a rundown for the two models for the equinoxes and solstices for my QTH for this year, with all times referenced to UTC:
|
Date |
Sunrise |
Sunset |
Noon |
|||
|
Old |
New |
Old |
New |
Old |
New |
|
|
3/20/2022 |
10:45:32 |
10:45:26 |
22:54:27 |
22:53:57 |
16:49:38 |
16:49:41 |
|
6/21/2022 |
09:08:11 |
9:08:13 |
00:20:00 |
0:19:53 |
16:44:06 |
16:44:03 |
|
9/20/2022 |
10:27:40 |
10:27:44 |
22:42:54 |
22:43:40 |
16:35:38 |
16:35:42 |
|
12/21/2022 |
12:05:38 |
12:05:38 |
21:15:12 |
21:15:00 |
16:40:25 |
16:40:19 |
For what it’s worth, Simon’s World Map provides sunrise/sunset times that closely align with the NOAA models, at least for today’s date.
I hope everyone has a great day.
Ken
From: Simon Brown
Sent: Monday, December 12, 2022 03:18 AM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
It depends how the other software calculates sunrise / sunset. I round to the nearest minute or 10 seconds. I believe my calculation to be correct.
--
Thanks for that information, Simon. The ‘Astronomical Algorithms’ by Jean Meeus were referenced in the older NOAA solar calculations, as well as in their newer calculations. It looks as though the first edition came out in 1991, and the second in 1999.
Thanks again for including such valuable information into such a useful program as World Map.
Merry Christmas, Simon.
Ken
Sent: Monday, December 12, 2022 01:03 PM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
If you start programming Sun / Moon / Planets then AA+ v2.47 (naughter.com) is a very good place to start.
AA+ is a C++ implementation for the algorithms as presented in the book "Astronomical Algorithms" by Jean Meeus. Source code is provided with the book, but it includes (IMHO) a restrictive license, as well as not having been updated for the 2nd revision of the book which includes new and interesting chapters, on areas such as the Moons of Saturn and the Moslem and Jewish Calendars. To make the most of my code, you will really need a copy of the book. This can be purchased from Amazon or directly from the publishers Sky & Telescope formerly Willman-Bell.
Example areas covered include the positions of the planets, comets, minor planets and the Moon, calculation of times of Rising, Setting and Transit, calculation of times of Equinoxes and Solstices plus calculation of the positions of the moons of Jupiter and Saturn as well as many other algorithms presented in the book. This is one of the biggest frameworks I have ever developed and includes 415+ thousand lines of code!
Simon Brown
Which sunset are you interested in? Geometric, civil, astronomical, yatta yatta
{^_^}
Hello all,
Simon is right on the mark. Different software programs can use different equations for calculating the time of sunrise and sunset. If the equations are scientifically sound and properly vetted, there can still be slight differences between the results they produce, certainly enough to result in differences of tens of seconds, or perhaps even up to and over 1 minute. I don’t think any of the equations are necessarily “wrong”, but some might be more “correct” than others. On top of that, the equations usually apply to a given astronomical epoch, so can produce slightly different results if you’re on the cusp of an epoch. In addition, some equations may be based on slightly different assumptions regarding polar and equatorial diameters of the Earth, and different assumptions regarding the curvature of the Earth surface. Then you can factor in assumptions of the precise orbital position of the Earth around the Sun for a given date. All of these factors can affect the precise estimation of sunrise/sunset. Again, the results may be very precise relative to the equation used, but might not agree exactly with another program that uses different equations which use slightly different constants and factors.
For example, here in the USA one can go to the NOAA (National Oceanographic and Atmospheric Administration) website and drill down to find data for sunrise/sunset for a specific location. I would trust that the equations used by NOAA are ‘correct’. Same with some of the propagation websites. There is useful program/app called LunaSolCal that can also calculate sun and moon data for a specific location, which I find to be very good and useful. If one checks a weather forecast program (Weather.com, Accuweather.com, WeatherBug, etc.), these will also display sunrise/sunset for a given location, and are also likely using a vetted source for their data. And I have no doubt Simon used a commonly approved/vetted set of equations to calculate solar and lunar data, which may or may not be the same as those used by other “authorities”. It may produce slightly different results, but technically I don’t view the results as “wrong” (they are indeed correct based on the mathematics of the equation used).
Also another thing to look out for. Many of the web-based sources of sunrise/sunset data (e.g. local weather forecasts) will display the data for a closest city in their database, as opposed to your actual QTH. That may be tens of miles distant east/west or north/south from the actual QTH where the user is located. That 10 or 15 or 20 miles can easily account for a difference in ±1 minute. Since Simon’s World Map uses a precise QTH based on whatever you provide in the setup, it should produce accurate sunrise/sunset times that are correct for the mathematics inherent in the equations Simon chose to use. Will Simon’s values agree exactly with a NOAA prediction for the same latitude/longitude? Maybe, maybe not, but I suspect the results would agree within ±1 minute.
Everyone have a great day.
Ken
From: Simon Brown
Sent: Monday, December 12, 2022 03:18 AM
To: main@SDR-Radio.groups.io
Subject: Re: [SDR-Radio] Simon's Worldmap - A question on sunset countdown time
It depends how the other software calculates sunrise / sunset. I round to the nearest minute or 10 seconds. I believe my calculation to be correct.