Navigational stars in Silent Hunter 5

Originally Posted by derstosstrupp

Huge thank you to Alpheratz and the creator of the NewSextant mod! I have always wanted a way to do celnav in a sim, and now all the pieces are there to do it. After seeing the post about the acceleration and deceleration of the celestial sphere, I was trying to figure it out in international TWOS - this last video and this information came at just the right time because my solution was inelegant at best.

I know sometimes when posting things, we are never quite sure how many people we are reaching or if anybody cares. So please know that myself and at least a couple others I know will greatly appreciate this effort. And I certainly hope that other players who might not as yet be familiar with celnav give this a try, it’s a very rich experience, and it’s really not hard (basic arithmetic is all it has to be).

Ritterkreuz verdient!

And noted for this year’s best of subsim vote when the time comes…

Thank you, derstosstrupp, for evaluating my work! I am glad that this attracts the attention of an authoritative and competent audience, which is interested in the fullest possible disclosure of the potential of our favorite simulator.

[derstosstrupp]

I have noticed that sometimes the NewSextant mod thinks a body is in a different spot than it appears to be. Typically for a noon latitude sight I don’t see UT show up after putting the object in. If I drop down a degree or so, sure enough, there it is. Now of course I don’t need time for a latitude sight, but it’s just odd. I have run into the same trying to shoot some stars, not very many, but sometimes I run into an issue where UT doesn’t show.

Regarding the sun, when I am doing a running fix for example, if I take, say, and afternoon sun line, for which I get UT displayed, when I use real almanac declination I get very close. However, when I use real declinations for a noon latitude sight (where UT doesn’t show up as mentioned before), I get very close using Alpheratz’s sun declinations. So there is definitely some disconnect perhaps in NewSextant. Not a big deal, because we can use “two sets of books”. But definitely something to investigate.

In short, if I can see UT show up in the display, I get very close using real declinations, but if UT doesn’t show, Alpheratz’s data seems to be closer. I have had a friend test this as well in parallel.

[vdr1981]

Originally Posted by derstosstrupp

I have noticed that sometimes the NewSextant mod thinks a body is in a different spot than it appears to be. Typically for a noon latitude sight I don’t see UT show up after putting the object in. If I drop down a degree or so, sure enough, there it is. Now of course I don’t need time for a latitude sight, but it’s just odd. I have run into the same trying to shoot some stars, not very many, but sometimes I run into an issue where UT doesn’t show.

Are you talking about campaign or single missions?
I noticed the same but only in some single missions. In campaign, it seems that new sextant works normally. Tested in Black Pit campaign, mid Atlantic, and UT is shown up to ~60deg of altitude...

[derstosstrupp]

This is in campaign. And it’s really only at high noon, and maybe with some selected stars, but not very many. I should also point out this is on TWOS version .21. I am waiting for my campaign transition to upgrade to the latest. My friend is on .22 though with the same issue.

[vdr1981]

Originally Posted by derstosstrupp

This is in campaign. And it’s really only at high noon, and maybe with some selected stars, but not very many. I should also point out this is on TWOS version .21. I am waiting for my campaign transition to upgrade to the latest. My friend is on .22 though with the same issue.

Ok, thank you for your report...


Any new strange things in the campaign which you previously didn't notice, without New Sextant?

Originally Posted by derstosstrupp

I have noticed that sometimes the NewSextant mod thinks a body is in a different spot than it appears to be. Typically for a noon latitude sight I don’t see UT show up after putting the object in. If I drop down a degree or so, sure enough, there it is. Now of course I don’t need time for a latitude sight, but it’s just odd. I have run into the same trying to shoot some stars, not very many, but sometimes I run into an issue where UT doesn’t show.

Regarding the sun, when I am doing a running fix for example, if I take, say, and afternoon sun line, for which I get UT displayed, when I use real almanac declination I get very close. However, when I use real declinations for a noon latitude sight (where UT doesn’t show up as mentioned before), I get very close using Alpheratz’s sun declinations. So there is definitely some disconnect perhaps in NewSextant. Not a big deal, because we can use “two sets of books”. But definitely something to investigate.

In short, if I can see UT show up in the display, I get very close using real declinations, but if UT doesn’t show, Alpheratz’s data seems to be closer. I have had a friend test this as well in parallel.

Thank you, you have raised a very important issue. I know what's the matter here. This problem you are talking about is not a bug in the mod. The reason lies in celestial mechanics. There is only one possible calculation method for time correction, and this method is implemented in the New sextant mod:
There is a formula that describes the change in altitude within 10 seconds of time:

Delta h per 10 sec = 2.5 arc min * cos (Latitude) * sin (Azimuth)

The New sextant mod calculation algorithm calculates the altitude for a celestial body and compares it with the Measured altitude.

Delta h = |Calculated altitude - Measured altitude|

Time correction is performed based on the above two formulas.

Now about celestial mechanics. After the rise of a celestial body, it increases its altitude relatively quickly. The rate of increase in altitude decreases as you approach the upper culmination. Immediately before upper culmination, at the moment of upper culmination, and immediately after upper culmination, there is practically no change in altitude. After upper culmination, altitude decreases very slowly at first, then it happens faster, the rate of altitude decrease is maximum before the celestial body sets over the horizon. The most accurate time correction is possible when the Azimuth of the celestial body = 90 or 270: this corresponds to the largest altitude increment per unit time. The possibilities for time correction are almost exhausted when you observe the upper or lower culmination of a celestial body, when there are practically no changes in altitude (Azimuth = 0 or 180).

And now the most interesting thing is why UT disappears. Even if you follow all the recommendations for working with the sextant, which I talked about in my videos, you can achieve an accuracy in measuring altitude at the level of + -0.5 ... 1.0 arc minutes. The upper culmination moment corresponds to the maximum altitude that a celestial body can reach at a given latitude. When at the moment of upper culmination you measure Altitude with an error towards overestimation of the result, let it be at least the smallest error + 0.1 arc minutes, then for the New sextant mod calculation algorithm you get an altitude that a celestial body can never reach - based on the measured altitude, the time cannot be corrected, so UT does not appear.

Everything I talked about above concerns both the measurement of stars and the Sun.
Now, for the Sun alone. I have to manually correct over 2500 lines of the Sun.txt file in the New sextant mod in order for this mod to be usable for time-corrected measurements of the Sun. When taking measurements of the sun, you need to make sure that the SEXTANTSUNDIAMETER (String # 53 of "page layout.py") matches the size of the sun in the mod being used. For example, the new test build TWoS 2.2.23 should have SEXTANTSUNDIAMETER = 2.18666666667, but the Sun.txt file has not been redesigned yet, it is premature to talk about it. When measuring the Sun, when in the sextant interface field you enter the value "0" to indicate the celestial body being measured, the sextant "knows" that you are measuring the lower limb of the solar disk and the sextant reading is the altitude of the solar disc center, taking into account the influence of dip + refraction. that is, you don't need to add sun semidiameter to the measured value.

If you want to check the relevance of the declination of a celestial body (remember the maximum accuracy of a single measurement is + -0.5 ... 1.0 arc minutes, and in the case of measuring the Sun due to blurred contours of the solar disk, the error can be 2 - 3 arc minutes), regardless work of New sextant mod and time correction, then you should take measurements without entering the number of the celestial body in the sextant interface field. The sextant reading in this case will represent the actual altitude (without the influence of dip + refraction), which you can use directly to calculate declination when measuring the altitude of upper culmination at a latitude known to you. When checking the declination of the Sun, you should also not enter "0" in the sextant interface field. Add Sun semidiameter to the sextant readings (1 deg 5.6 arc minutes in new test build 2.2.23) and calculate the declination.

[derstosstrupp]

This makes sense. In the meantime, for noon latitude sights (where the problem is most pronounced) your declination numbers for the sun have yielded nearly spot-on results (as you mention, to compensate for the too-high altitude and thus too-low zenith distance). In game I normally get a morning twilight 3-star fix, and then a running fix during the day using a noon latitude sight advanced to another afternoon sun line (normally 3 hours later for a good azimuth change). For the afternoon sun line I always have UT displayed anyway so I can use actual declination (so no issues).

So it’s 2 sets of books for now and that’s OK.

I have also adjusted the diameter of the sun in the mod, mine corresponds to your value of just over a degree.

Thanks for posting that formula also. I will take a closer look at that.

Originally Posted by derstosstrupp

For the afternoon sun line I always have UT displayed anyway so I can use actual declination (so no issues).

So it’s 2 sets of books for now and that’s OK.

You cannot rely on UT and draw conclusions on this basis (including regarding the correctness of declination) if you read UT when measuring the Sun, because there are 2 precisely established facts:
1. The declination of the Sun SH5 does not correspond to the actual declination of the Sun (you can check on your mod with the checkpoints on March 20, June 22, September 23, December 22).
2. The time correction is performed on the basis of the actual declination of the Sun (see file Sun.txt), and therefore gives a deliberately incorrect result.
The only exceptions are measurements taken around March 20 and September 23.

Originally Posted by derstosstrupp

In game I normally get a morning twilight 3-star fix, and then a running fix during the day using a noon latitude sight advanced to another afternoon sun line (normally 3 hours later for a good azimuth change).

If you test the New Sextant mod with this method, keep in mind that there are limitations due to the fact that WORLD = MAP in SH5. This imposes restrictions to advance one line of position to another line of position made after a long time (observations with baseline shift, "Sonnenbeobachtungen mit Versegelung"). In fact in SH5 you only have 2 options.

1. Between the measurements of the first and second altitudes your Heading = 0 or 180, then this formula works to transform the first taken altitude and refer it to the zenith of the second observation:

delta h = Speed ​​[knots] * 0.9265 * time [hours] * cos (Azimuth - Heading)

2. Between the measurement of the first and second altitudes your Heading = 90 or 270, then the corrected first formula works:

delta h = Speed ​​[knots] * 0.9265 * time [hours] * cos (Azimuth - Heading) * cos (latitude)

If Heading is different from 0, 180, 90, 270, then it is impossible to calculate "delta h" using a calculator or tables.

Thank you very much, derstosstrupp, for your prompt assistance in correcting sun declination in New Sextant mod. My preliminary testing revealed no errors. Link to the sun.txt file, which should replace the corresponding New Sextant mod file:

https://drive.google.com/file/d/1cw6...ew?usp=sharing
I propose to test this on measurements of the sun. I will continue testing as well.

NOTA BENE! My guess turned out to be wrong.

Originally Posted by Alpheratz

When measuring the Sun, when in the sextant interface field you enter the value "0" to indicate the celestial body being measured, the sextant "knows" that you are measuring the lower limb of the solar disk and the sextant reading is the altitude of the solar disc center, taking into account the influence of dip + refraction. that is, you don't need to add sun semidiameter to the measured value.

When you measure the lower limb of solar disc altitude, then New Sextant readings represent the sun's altitude reduced by the actual (historical) sun semidiameter. You need to add the historical sun semidiameter to the measured sun lower limb altitude. I will rework my Sun Almanac to include a historical sun semidiameter.

[derstosstrupp]

Originally Posted by Alpheratz

Thank you very much, derstosstrupp, for your prompt assistance in correcting sun declination in New Sextant mod. My preliminary testing revealed no errors. Link to the sun.txt file, which should replace the corresponding New Sextant mod file:

https://drive.google.com/file/d/1cw6...ew?usp=sharing
I propose to test this on measurements of the sun. I will continue testing as well.

NOTA BENE! My guess turned out to be wrong.

When you measure the lower limb of solar disc altitude, then New Sextant readings represent the sun's altitude reduced by the actual (historical) sun semidiameter. You need to add the historical sun semidiameter to the measured sun lower limb altitude. I will rework my Sun Almanac to include a historical sun semidiameter.

You are very welcome! I will test tonight as well.

[derstosstrupp]

Not sure how many others are doing full celnav out there, but for those who are, please share your routine (your “day’s work in navigation”). Here’s mine below. For sight reduction I use various tabular methods (S-table, F-Tafel, or Fulst’s Nautical Tables) or a calculator if not feeling up to crunching logarithms.

1. Morning twilight 2-star fix.

2. Calculate LAN for the approximate longitude I will be around LAN.

3. Take a sun sight about 3 hours before LAN for a sun line.

4. Take a latitude sight at LAN and advance the morning sun line for a running fix.

5. Take a sun sight about 3 hours after LAN. I won’t typically reduce this unless I can’t get evening stars or am otherwise diverted (by a target sighting etc). In that case I’ll advance the LAN latitude line for another running fix.

6. Calculate time of sundown and plan on being on the bridge shortly thereafter.

7. Evening twilight 2-star fix.

8. Calculate time of sunup and plan to be on the bridge about 30 minutes prior.

9. Hit the sack - I’ve earned it.

For my dead reckoning, I keep a log in a notebook, making an entry each time I change course, speed, save and exit, or get a fix.

The one concession I make is I pretend every target I sink radios its position and that it’s accurate. It’s my reward for sinking things.

Please share your approach!

I showed my CelNav routine in the video "Silent Hunter 5 Wolves of Steel: what is Real navigation in 2021" and in general it is similar to what derstosstrupp does:

https://www.subsim.com/radioroom/sho...62521&page=240

Of the two stars, one is usually Polaris to save time.

For calculations I use either F-Tafel or my excel spreadsheet "SH5 celestial navigation Sun, Moon, 58 stars" (this is an updated version of my excel spreadsheet "SH5 celestial navigation" included in C: \ Ubisoft \ Silent Hunter 5 \ The Wolves of Steel - Documentation \ Real Navigation \ RealNav - almanacs and tables by Alpheratz), which can be downloaded from the link in my signature.

I recently got acquainted with Fulst's Nautical Tables, which seemed to me quite easy to use (many thanks to derstosstrupp for explaining and showing me how to use them).

[SnipersHunter]

I am currently starting to lern celnav thanks to Alpheratz excellent series. Can not wait to do a full patrol without needing my navigator.

Originally Posted by SnipersHunter

I am currently starting to lern celnav thanks to Alpheratz excellent series. Can not wait to do a full patrol without needing my navigator.

I am delighted to hear that and you can count on my help if you need it.