Meadville Space Center
Welcome, Guest. Please login or register.
December 09, 2019, 05:22:12 PM

Login with username, password and session length
Search:     Advanced search
Project Apollo - NASSP 6.4.3 released!
http://nassp.sf.net
25068 Posts in 2094 Topics by 2266 Members
Latest Member: twa517
* Home Help Search Login Register
+  Meadville Space Center
|-+  Project Apollo - NASSP
| |-+  Project Apollo - NASSP News & Discussion
| | |-+  Support & Bugs (Moderators: movieman, Tschachim, Swatch, lassombra)
| | | |-+  P23 Cislunar Navigation
« previous next »
Pages: [1] 2 Print
Author Topic: P23 Cislunar Navigation  (Read 7192 times)
jacarpe3
Jr. Member
**
Posts: 2


View Profile Email
« on: October 21, 2011, 01:41:13 PM »

Hello everyone.  I'm a newbie to Project Apollo and am thoroughly impressed with the level of detail and commitment shown by its developers.  I am working through the vAGC Apollo 7 scenario and have hit a snag at the P23 Star/Horizon Sighting (GET +123:00).  The procedure as outlined in the flightplan in the "AGC" page (doc\Project Apollo - NASSP\Flightplans\Apollo 7\Apollo 7 Flightplan vAGC.xls) tells you to mark the star when the display is flashing V59.  It then has you mark the horizon when the display is flashing V51.  Everytime I go through this procedure precisely and I get wildly incorrect SV change in velocity and position numbers.  I literally follow the instructions word for word.  I looked at the G&C Checklist (doc/Project Apollo - NASSP\Word Checklists\G&C Checklist) and the P23 procedure there says that a flashing V59 means "request Optic Calibration".  Flashing V51 is the actual mark from what I can tell.  Also, when I use the G&C Checklist it mentions to use V to activate the split line of sight.  when I do this, the only way I can get the two views to concide during the calibration (flashing V59) is when my trunnion and shaft angles are zero.  Once I move the sextant to the horizon for the mark the LLOS and SLOS are totally out of whack again and I can't go on with the procedure as I understand it.  I'm so lost!  The two procedures I've mentioned above seem totally different.  The one in the excel flightplan doesn't mention the split LOS, calibration (flashing V59).  I've been working and researching this for 2 days now and it's really irritating  Bangs Head 2  If I somehow missed the answer to this question somewhere in the forum I apologize but I have looked for days!  Literally!   HELP!!!

Quote
...and I can confirm that star/earth landmark sighting with P23 works neatly (4 balls one and all balls), too. (http://www.ibiblio.org/mscorbit/mscforum/index.php?topic=2715.0)

How?  I get 2-3 balls at best.

Thanks again in advance.

PS - here is my scn file

* Apollo 7 vAGC P23 Cislunar Navigation.scn (153.94 KB - downloaded 177 times.)
Logged
Apathy Man
Project Team Member
Full Member
****
Posts: 17



View Profile Email
« Reply #1 on: October 21, 2011, 01:50:49 PM »

I'll be working on rewriting the flightplan files in the near future. Perhaps there had been changes committed between now and when my brother had committed them that may not permit P23 to work. I'll be sure to focus on it and develop the procedure effectively when I finalize testing for the "meat" of the mission. Rendezvous has me snagged up for the time being.

From reading these flightplan files, it seemed my brother had found ways or sometimes work-arounds to get things as historic as possible. The difficult thing I'm faced with is trying to understand the ways he used to achieve these objectives, which at times seem pretty vague to me.
Logged

"Whoopee!" Pete Conrad's first word upon stepping on the moon.
eddievhfan1984
Hero Member
*****
Posts: 737



View Profile
« Reply #2 on: October 21, 2011, 06:00:29 PM »

Stick with the version from the G&C checklist, it's more accurate. And you won't be asked to calibrate until after you've set the desired star/landmark/horizon settings. Make sure the spacecraft is stable from the auto manuever, and only use telescope pitch (shaft?) to line up the star with the horizon.

I hope this doesn't seem patronizing.
Logged
meik84
Project Team Member
Sr. Member
****
Posts: 454



View Profile
« Reply #3 on: October 21, 2011, 06:19:50 PM »

Quote
I am working through the vAGC Apollo 7 scenario and have hit a snag at the P23 Star/Horizon Sighting (GET +123:00).
Uuuh. Doing P23 in earth orbit is not a very precise thing. Earth's atmosphere seems (optically) thicker, so that you can't tell where the 'hard' part starts and the 'soft' part ends. This makes marking difficult and thus P23 isn't very precise. When you get further away from earth, this effect decreases and so your markings will get better. This is no flaw of orbiter. Apollo 7 had the same problem, Apollo 8 was the first that demonstrated the very high accuracy of P23 (and brought Jim Lovell the nickname 'the man with the golden hands').
Quote
doc\Project Apollo - NASSP\Flightplans\Apollo 7\Apollo 7 Flightplan vAGC.xls
Forget those. The word checklist is technically right, but not very easy to understand. I'll talk you through it, if you don't mind. Wink
Assumptions:
- DAP on & initialized
- W-Matrix initialized
- IMU on & aligned
- state vector updated (recommended)

G/N PWR OPTICS– on
OPT TELTRUN – SLAVE to SXT [optional; I for myself place it to 0°, so that the SCT coincides with the LLOS]
OPT SPEED - HI
OPT MODE – ZERO (15 sec)
OPT MODE – CMC
Star markers - activated, Apollo star markers selected   {F9}

V37E23E

F 05 70

R1 000XX  STAR ID
R2 00CDE  LMK ID
R3 00CF0  HOR ID

The infamous Noun 70 (note that all this is octal -leading 0's can be omitted when entering new data). Here we tell the CMC what we are going to do.
R1 selects the star we want to mark on.
R2 is used when we want to mark with a star and a landmark (Apollo 8 did that on one occasion - I don't know if 7 did.). C selects the body the LMK is on: 1 for earth, 2 for moon, 0 for 'no LMK measurement'. D and E select the LMK. 00 for 'astronaut defined LMK coordinates' (the CMC will ask you later for the precise coordinates), everything else for predefined (i.e. stored in CMC's fixed memory) coordinates. As we don't want to mark on a LMK, log in all (or, if you are a 'CMC crack', one) 0.
R3 is used when we want to mark between a star and moon's or earth's horizon. C is analog to R2 (0 means 'no horizon measurement'), F is the 'horizon option'. Horizon option? Well, look at the full moon at night: it's a round disc. Now, imagine you want to measure the distance between the moon's edge and a star: you could either use the edge (or horizon) that lies next to the star (the 'near horizon') or the one that lies 'on the other side' of the moon (the 'far horizon'). The CMC needs to know which horizon we use, so that it can determine the correct angle between the star and the horizon. 1 is for 'near horizon', 2 for 'far horizon'.

V25E to load new data. I recommend R2= 00000 , R3 = 00110. R1 is left to your choice -you should pick a star which is not too near and not too far from the horizon; the CMC will check if it is visible and in reach of the optics by itself.

PRO to proceed.

F 50 25

R1 00202
R2 & R3 blank

'Please perform auto maneuver'; the CMC has computed an attitude that will point the LLOS at (or at least near) the horizon.

ENTR if you want to maneuver by yourself

SC CONT - CMC
CMC MODE - AUTO
PRO
for letting the CMC do that job.

F 50 18

The attitude in roll, pitch and jaw. PRO to start auto maneuver.

06 18

Maneuver in progress. When it's done,

F 50 18 will reappear.

ENTR to

F 59

'Request optics calibration'. The real optics had some sort of 'bias' in it: they 'expanded' when exposed to the sun and 'shrunk' when they faced cold space. To knock that out, the LLOS was pointed on a star and the same star was superimposed with the SLOS. The CMC now knew the bias and could incorporate it in its calculations. The NASSP optics don't have any bias, so we just put OPT MODE to ZERO and mark (keyboard key Q) once.

F 06 87

R2 +00000

The trunnion bias. All 0's, as our optics are zeroed.
OPT MODE - CMC
PRO to

06 92

R1 XXXXX
R2 XXXXX

Shaft and trunnion auto optics angles. The CMC tries to keep the optics pointed near (a 'direct hit' is nearly impossible) the star. As soon as the OPT MODE switch is moved out of the 'CMC',

F 51

will occur. Now the real work begins: Activate the split screen and try to place the star on the horizon. One very important thing is that you've got to place it on the point of the horizon that lies most next to the star -normally you'll have to maneuver to point the LLOS appropriately. I recommend switching to SCS MIN IMP (i.e. SC CONT - SCS, MANUAL ATTITUDE (3) - MIN IMP), so that you have the right hand for the spacecraft and the left for the optics. To get a rough direction where to point at use the SCT at 0° trunnion, impose the yellow line on the star and maneuver down so that the circle points at the horizon. Don't make wild maneuvers, fire single bursts and wait for the spacecrafts reactions. When the circle is on the horizon, switch to the SXT, put the horizon in the middle of the cross, place the star on it and mark to get

F 50 25

R1 00016

'Please terminate marking'. Key PRO.

F 05 71

The same as F 05 70 above. The CMC asks you to verify the data. Key PRO.

F 06 49

R1 DELTA R in .1 nm
R2 DELTA V in .1 fps
R3 blank

Finally! From our mark, the CMC has calculated a new state vector. He compares it with the actual state vector and shows the differences in position ('delta r') and velocity ('delta v'). If our mark was good (and the state vector not to old), this should be small -0000X for both is 'Lovell-like', 0002X not bad. Anything above 00100 is bad and should be discarded by keying V32E. You will return to N70 to make a new mark. If everything is good, key PRO. The CMC updates the state vector and comes with

F 37

to remind you to select a new program.

Don't worry if your first marks are bad - that's a thing of training, experience and a lot of patience. The scenario I posted in the thread you mentioned is a neat sandbox for this... Wink
Logged
jalexb88
Sr. Member
****
Posts: 328


View Profile
« Reply #4 on: June 03, 2014, 11:29:57 AM »

Wow following the vAGC apollo 8 flight plan I was always getting 00000, but the correct procedure is in the G & C checklist. I knew I wasnt that good lol!  One question: In the real Apollo 8 flight plan it shows how Jim Lovell had to make 2 or 3 different "sets" of several markings each. How do we do this? Is it a question of redoing P23 over again 2 or 3 times? When I do it I seem to only be able to mark once but the flight plan would suggest that somehow he did multiple markings in one go.

Alex
Logged
eddievhfan1984
Hero Member
*****
Posts: 737



View Profile
« Reply #5 on: June 03, 2014, 11:45:16 AM »

That's what V32 was for. After they marked and got the data, they'd then recycle the program and do it again.
Logged
jalexb88
Sr. Member
****
Posts: 328


View Profile
« Reply #6 on: June 09, 2014, 08:40:06 PM »

I've now been messing around with P22 orbital nav in lunar orbit on Apollo 8 and for the most part in works great! On the real Apollo 8 mission they used 3 "control points" and a pseudo landing site they called "B1" and according to the real flight plan they used the optics to get a lat long reading for each control points and the landing site. This gave them updated more accurate coordinates from the estimates that were based on old photos from previous unmanned missions.

According to some transcripts on the Apollo 8 Flight Journal https://history.nasa.gov/ap08fj/ The control points did not seem to have an assigned number in the AGC so they i'm guessing had to select a "known landmark" which in 05 70 is R2: 10000, and then enter the coordinates manually. However the landing site seemed to have a code, presumably 01, so R2 was 10001 and auto optics would track it. Looking at the pad loads excel file I see RLS lunar landing site 2025-2032 and I believe this is where "B1" point was loaded. Anyone know how one could convert the LAT LONG numbers for this point from the flight plan into EMEM format?

Alex
« Last Edit: June 09, 2014, 08:43:11 PM by jalexb88 » Logged
meik84
Project Team Member
Sr. Member
****
Posts: 454



View Profile
« Reply #7 on: June 10, 2014, 04:05:34 AM »

The problem is that the flightplan gives the coordinates in latitude, longitude (/2) and altitude, but CMC stores it as a 3D-vector. For the CMC, that was no problem: it had a special subroutine (the latitude-longitude subroutine) to convert between both systems easily. This means for us that we'd have to emulate this routine. From first sight it looks quite easy, just some vector/matrix computations, but I'd really go ahead with LVDC and SECS...
Logged
jalexb88
Sr. Member
****
Posts: 328


View Profile
« Reply #8 on: June 10, 2014, 08:28:25 AM »

Maybe If we can somehow save the B1 coordinates within the vAGC and extract the EMEMs from the saved scenario state? Anyhow I agree there are more pressing issues to work out before this and meanwhile all one has to do is enter the coordinates manually in P22.

Alex
Logged
meik84
Project Team Member
Sr. Member
****
Posts: 454



View Profile
« Reply #9 on: June 10, 2014, 11:37:51 AM »

Quote
Maybe If we can somehow save the B1 coordinates within the vAGC and extract the EMEMs from the saved scenario state?
Hm. When you hit PRO on the flashing V06N89 that follows after the flashing V06N49, the CMC stores the coordinates shown in the mentioned adresses. Quicksave and you can extract the EMEMs. However, you'd have to mark on the landing site first, so that would wouldn't be the flight plan coordinates, but those revised by the CMC.
Logged
jalexb88
Sr. Member
****
Posts: 328


View Profile
« Reply #10 on: November 18, 2014, 12:02:00 AM »

I've finally gotten the CMC to calculate the coordinates for the pseudo landing site and have extracted the EMEMs for it.

  EMEM2025 256
  EMEM2026 61325
  EMEM2027 171
  EMEM2030 25022
  EMEM2031 12
  EMEM2032 74063

That will store B-1 in the CMC at  Long/2 +17512,  Lat +02675. These are the flight plan coordinates since I cannot find any documents indicating where B-1 was finally located so I went with that.

I am attaching a scenario and ground landmarks marker file which starts at revolution 7 of the Apollo 8 mission with control points 2 and 3, followed by the pseudo landing site B-1 just before sunset. This provides for a good exercise of P22 in lunar orbit and it works quite well by following the P22 GNC checklist. When you see B-1 on the horizon, run a P22. At 05 70 type in 10001, proceed and the optics should lock onto it. (B-1 appears just before sunset so you must wait a few minutes to reach it under 10x time accel)

Place Apollo Navigation Landmarks.mkr in \Config\Moon\Marker and activate surface landmarks

For anyone that would like to practice sightings on CP-2 and CP-3 as well as B-1, here is how to get the auto optics to find them:

Run my scenario, wait until the desired CP is well within the field of view then

V37E 22E, PRO to 05 70

05 70: Enter 10000 in R2, PRO     (If you are doing landmark B-1 then enter 10001, its position is known to the CMC so it will bypass 06 89 below.)

06 89: Enter the Lat, Long/2, Alt      (Long is divided by 2 from the actual value to be able to fit in the 5 digit dsky register)

                CP1           CP2          
LAT        -09675       -09150    
LONG/2 +81656      +01935
ALT       +00000     +00000

OPTICS MODE - AUTO
PRO

The optics will lock on the traget then switch to OPTICS MODE - MAN and make 5 marks leaving a bit of time between each mark.

50 20:  0016 is displayed when sufficient marks have been made. Hit PRO

05 71: Enter 20000 in R2. This tells the CMC we want it to compute new coordinates for the landmark. Hit PRO

06 49: New state vector from our markings, should be close to all balls. Hit PRO

06 89: These are the new coordinates calculated for the landmark, it should be fairly close to what was entered previously. Do not hit PRO here as it will update the landing site in the CMC.

Hit V34E to terminate the program and run it again on the next landmark.


Alex




* Apollo 8 - P22 Lunar.zip (21.76 KB - downloaded 153 times.)
« Last Edit: November 18, 2014, 03:41:13 PM by jalexb88 » Logged
indy91
Hero Member
*****
Posts: 1316


View Profile
« Reply #11 on: February 05, 2015, 08:11:16 PM »

This isn't exactly the right thread, but you guys started to talk about P22 in a P23 thread, so it's your fault Tongue

So I have finally reached the point during flying the Apollo 8 mission, where I have to do the landmark tracking. Jalexb88's post above has been very helpful for that, so thank you! I noticed, that the way you got the ENEMs for the B-1 landmark seems rather imprecise. So I used a little bit of math, that I found in the GSOP for Colossus, page 5.5-21, to calculate the ENEMs from the longitude and latitutde. I used the flight plan values and for the moon radius the Orbiter moon radius instead of the AGC moon radius (there is not much of a difference). The resulting numbers were:

EMEM2025 255
EMEM2026 21676
EMEM2027 171
EMEM2030 24251
EMEM2031 11
EMEM2032 34673

I tested it with the scenario in the post above and the optics seem to track the landmark a little bit closer. Not 100% sure yet my values are actually better. Was the B-1 location padloaded? If yes, then we should probably go for the most accurate values we can get Happy
« Last Edit: February 05, 2015, 08:18:17 PM by indy91 » Logged
jalexb88
Sr. Member
****
Posts: 328


View Profile
« Reply #12 on: May 11, 2015, 02:57:02 PM »

Okay I've just completed another Apollo 8 attempt with indy91's latest MFD updates which are nothing short of amazing. Unrelated to that I've been messing around with P23 cislunar navigation, and have found a snag: It does not do what its supposed to do very well, which is update the state vector. In fact, I find it makes the state vector worse.

 I followed the procedures earlier in this thread outlined by meik84, everything works great, Ive updated the w-matrix with +00094, +00057 +00003 and proceeded with the markings with nearly all balls on F 06 49. I then checked to see what this actually did to my state vector, I updated the "other" slot with a ground SV for comparison using V83 to see how close my "this" slot (P23) state vector is to the "other" slot (ground) state vector. With every pass through P23 this distance in V83 increases more and more, the state vectors do not converge as they should.

Ive tried various combinations of stars, far horizon, near horizon, moon, etc. and many, many times over a very long portion of trans-lunar coasting. Something isnt adding up here, maybe the earth / moon horizon isnt where the CMC expects it to be with reference to the stars or the trunion angle is not read correctly, (which I doubt). Any ideas?
« Last Edit: May 11, 2015, 04:50:19 PM by jalexb88 » Logged
meik84
Project Team Member
Sr. Member
****
Posts: 454



View Profile
« Reply #13 on: May 11, 2015, 05:18:32 PM »

Well, for earth horizon/star measurements it is possible that (again) orbiter's spherical earth is the guilty part. CMC's idea of the horizon ellipse is based on Fischer's ellipsoid and not on a perfectly round earth. No idea how much difference this makes, though.
However, moon orizon/star measurements should work, as both CMC's and orbiter's moons are perfect spheres. I've tested that a long time ago and have the vague memory that it worked quite well, considering that finding the sub-stellar point is quite a try-and-error thing.
Most interestingly, earth landmark/star measurements worked perfectly from my memory, even though CMC works with Fischer's ellipsoid there again. Confused
Logged
indy91
Hero Member
*****
Posts: 1316


View Profile
« Reply #14 on: May 11, 2015, 05:53:11 PM »

Most interestingly, earth landmark/star measurements worked perfectly from my memory, even though CMC works with Fischer's ellipsoid there again. Confused

I think I can explain that. The radius of the spherical Earth is set to the actual launchpad radius, which means that at every latitude close to +/- 28.52° the Fischer ellipsoid has the same radius as the Orbiter Earth. Isla Tiburon, the landmark for P23, has a latitude of 28.876°N, which is pretty close.
Logged
Pages: [1] 2 Print 
« previous next »
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.10 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!