[13:14:07] NASSP Logging has been started by n7275 [13:14:09] hello [13:29:36] hey Matt [13:36:17] ah at least the Apollo 17 LGC padload agrees bit by bit with the IMU bias compensation [13:36:23] so the scaling and such is probably all good [13:37:00] I'll push that update. In a next step I would add input boxes for the GUI and mission specific numbers for the flown missions. [13:39:39] the Apollo 17 LM seems to have had a terrible PIPA looking at these large values :D [13:43:36] what was the acceptance criteria for these values? certainly above some value they would've just been rejected [13:55:07] hmm, haven't seen any numbers on that before [14:56:46] I will probably run all my numbers through your padload calculator before we even think about merging, just to double check. [14:58:52] yep, I am approaching it purely from a padload perspective. Then we will see if we have agreement about scaling and all that [15:03:57] I did try letting the Apollo 11 CMC drift for 24 hours with compensation. The misalignment at the end was so small I thought I was in the wrong branch [15:05:35] that's how it should be haha [15:52:30] I tried flying PDI with no compensation [15:52:50] it doesn't seem to like it very much [16:01:57] with Apollo 11 biases? [16:02:37] yes [16:03:48] around 2:50 into PDI the LM which had been in a stable attitude up to that point decides to do some flips [16:05:33] hmm [16:05:44] that doesn't really sound like a drift issue to be honest [16:06:56] no [16:07:24] alignment is only 0.03° off [16:13:16] I can't think why it would be doing this. I'm just using our saved scenerio. I should try in the Orbiter2016 branch [16:13:31] was the FDAI doing anything weird? [16:13:53] maybe one of your changes affects the IMU attitude given to the LGC somehow [16:15:33] FDAI looked normal [16:17:52] hmm, if I push the abort button, it recovers and gets me back into a normal looking post-abort orbit [16:22:40] so probably not a PIPA issue [16:25:09] yeah, I would suspect not [16:44:54] Orbiter2016 branch, around PDI+6 https://imgur.com/a/DaY596A [16:46:09] oh [16:46:56] hmm, there really hasn't been any change recently that could do this [16:47:02] could it be a build issue? [16:47:07] I will try it in a bit as well [16:49:31] It was holding the descent attitude then, then very slowly tipped backwards (pitch up) did a single flip and stabilized again. The same thing happened twice more, until fuel ran out [16:53:01] ok, rebuilding to check for myself [16:53:08] Apollo 11 Before PDI scenario, right? [16:53:51] yes [16:54:43] I am not all that convinced it will happen to me, but we will see :D [16:58:53] The likelyhood that I just did something dumb is statistically, at least, a significant posibility [17:05:14] morning! [17:05:46] mascons have gotten to your head (or rather your LM) [17:05:53] hey Mike :D [17:09:47] yeah, mascons: the real cause of space-madness [17:14:01] I'm at TIG+3min, all good so far [17:16:26] when I tried in the main branch it was around +6 [17:17:55] all good here at throttle down [17:18:30] weird [17:19:21] joystick connected? This is Orbiter Beta and Orbiter2016 branch, right? [17:19:31] anything with the mission file... [17:19:42] no local changes [17:19:48] ? [17:21:19] I do have a joystick connected. hmmmmmm speaking of biases, I need to check something [17:22:20] maybe you did what I did in Elder Scrolls Oblivion many years ago then :D [17:22:29] it took my connected joystick instead of my mouse [17:22:33] took me a while to figure out... [17:22:54] there was a slight drift all the time which confused me a lot more than mouse not working at all [17:23:04] "why am I walking sideways" [17:28:49] well it looks pretty well centered. maybe it is a "hero of Kvatch" problem though [17:37:56] and aside from hardover, the ACA input shouldn't really matter in PGNS Auto [17:45:28] yeah [17:45:34] it didn't fix the issue [17:45:44] and I missed something [17:46:31] the FDAI needles are deflecting rather significantly before the attitude departure [17:50:44] the needles, hmm [17:50:53] so the PGNS does want to do something [17:51:16] the IMU pitch is fine before that happens? I was sort of thinking the LGC gets bad IMU input somehow [17:52:54] could LR data be doing this? [17:53:45] no [17:53:52] would have messed up the state vector [17:58:53] it appears fine. the needle definitely moves first, then the ball as far as I can tell visually, follows spacecraft attitude. your IMU alignment debug tool confirms this [17:59:22] seems more like a state vector thing [18:01:22] and you definitely did a rebuild after switching branches [18:01:32] VS open with unsaved changes? That has caused me trouble one time [18:01:54] and the scenario is also stock [18:01:59] Git GUI shows no local changes [18:02:11] yeah definitely did a rebuild all and confirmed that everything built [18:02:17] no local changes [18:03:03] I can send you a scenario in a little bit saved right in the middle of the flip [18:03:28] I'm out picking up lunch right now, back in a few minutes [18:16:58] Maybe gravity gradient torque or radiation pressure or something weird like that [18:20:07] i sent it on discord [18:20:55] got it [18:26:38] hmm [18:26:51] the state vector seems pretty bad, but otherwise I can't tell much from it yet [18:41:51] have you tried it without LR yet? [18:46:55] not yet [18:47:21] didn't you have a flip at 2:50 though [18:47:23] before LR data [18:47:29] before the 180° yaw even [18:48:36] well yes [18:48:42] so that rules out LR data [18:49:36] accidentally merged something to the Orbiter2016 branch? [18:50:58] oh possible haha. let me check [18:51:34] it sounds like a OO sort of issue :D [18:51:39] Do we normally get RCS flring during descent? [18:51:46] yep, a bit [18:51:50] in yaw anyway [18:52:00] also, is our DPS animated? [18:52:09] but if the DPS gimballing isn't quick enough for the DAP then it fires RCS [18:52:11] DPS is animated [18:52:39] I'd need to check somedebug info, but I swear it's not doing much [18:53:23] it shouldn't [18:53:36] it's slow any should only really point along the longitudinal axis [18:53:42] full LM CG is still centered [18:53:50] slow and* [18:53:54] ahh [18:54:05] you can do a V48 if you want to see it move the 6° [18:54:20] in fact I have a scenario from you before DOI where I can show you that :D [18:55:53] you could watch PIPA pulses next attempt [18:55:55] N21 I think [18:55:58] haha, oh yeah [18:56:33] My Orbiter2016 branch exactly matches upstream/Orbiter2016 [18:57:33] I need one of those LGC telemetry printouts [19:02:21] maybe it's some Orbiter setting [19:02:51] What should I be seeing for PIPA pulses [19:02:59] I'm not sure actually haha [19:03:52] R3 is jumping between -0056X and -1160ish [19:04:23] reading data might clear these registers every 2 seconds [19:05:24] oo we pitched down this time [19:05:46] nothing remarkable in N21 beforehand [19:13:49] how is the Eagle? Landed? [19:14:00] oh [19:14:12] by "pitched down this time" you meant flipped around, but the other way [19:14:22] not "normal pitchover" [19:17:23] and the flipping is done by RCS? [19:17:30] or the DPS, can you tell at all? [19:17:39] or some magic external torque [19:42:14] seems to be a combination of DPS and RCS [19:53:52] I guess that indicates the LGC wants to point the LM somewhere else [19:53:59] not a DPS or RCS or IMU problem, right? [19:54:15] I still find it suspicious that you worked on an IMU topic and now get this issue [19:54:24] feels related [20:09:46] yeah [20:14:00] I had two more theories, but both of them are wrong according to your scenario [20:16:59] if this is "real" behavior, then it would have to do with too much/too little thrust or mass or so [20:17:40] LGC trying to compensate for it and eventually flipping around [20:33:18] the thrust command is doing some odd things [20:33:24] needle that is [20:36:05] interesting [20:36:16] if it's not at 100% all the time that could be the cause [20:37:13] hmm [20:37:17] the actual thrust would [20:37:29] thrust command points to the LGC being confused about something [20:40:04] yeah, thrust command looks like it's trying to throttle way down too early [20:40:51] this has to be a PIPA scale thing [20:42:07] which can't really be because you aren't in your branch [21:30:46] this defies logic. [21:31:37] I would go through all the Orbiter settings, if there is anything unusual [22:06:33] night! [22:47:02] n7275: do you know anything about the uplink side of things? or have you been only focusing on downlink? [22:49:07] I know less than I'd like, but I've tried to pay attention to both in the documents I've looked through [22:49:32] What are you interested in [23:00:00] There's a document I'm trying to find " Command Flow" or something similar [23:11:19] we're trying to get our UDL working and have some evidence that the sub-bit patterns shown in the system handbooks may not be correct [23:15:09] This may have some info: https://www.ibiblio.org/apollo/Documents/Apollo%20Command%20System%20-%20Ground%20Net%20Data%20Flow.pdf [23:20:33] hmm, maybe not [23:36:25] I'm pretty sure the sub-bits are added to the Univac 642s to the incoming command message [23:36:43] by the 642s [23:45:59] there's a better document, and now I can't find it. it's the one I always find when I'm looking for downwind stuff, but it covers mostly uplink [00:10:27] oh actually thewonderidiot, page 20 of that document [00:10:49] "typical sub bit patterns" [00:18:16] amazing [00:18:24] different from the systems handbook, and also different from the actual hardware :D [00:20:14] oh interesting [00:21:11] so I take it you figured out what they actually are? [00:23:50] yep, we just flipped our first UDL relay [00:25:35] for vehicle code: 1 = 11000, 0 = 00111; for the rest of the bits: 1 = 01011, 0 = 10100 [00:34:04] interesting that there's so much variety in something where there's only 32 possibilities [00:36:40] hahaha yeah [02:44:01] .tell indy91 Aparently my joystick is sending out some random commands, *and* in addition to that: https://github.com/orbiternassp/NASSP/pull/1030/commits/e9c78884657e89fcb956f668ec761e4faaa819a7 [02:48:53] oh man :D [02:49:13] that will definitely make a difference haha [03:17:56] yeah, I was looking at PGNS vs AGS VI and finially noticed that my acceleration was about double what it should be [15:50:07] hey [15:54:19] but that's a commit in your branch not in Orbiter2016 [17:15:38] hello [17:17:23] indy91 this was two seperate issues, one of them affecting everything, and one of them only affecting the IMU branch [17:17:35] yep, makes sense [17:17:44] annoying joystick :D [17:18:01] so it did what, caused hardover? [17:18:21] yep, random full deflection ACA input [17:19:01] I need a better joystick [17:20:49] I had one, and I misplaced it the last time I moved [17:21:03] my joystick is old, but thanks to force feedback there have never been any centering issues. The ACAs have fairly strict deadzone limititations, so most people have had some trouble, usually in yaw. [17:21:21] deadzone requirements I should say [17:21:26] so now I'm using an old RealFlight controller + Vesim [17:22:36] it has some weird features that I dont understand like (I believe, at least) biasing the analog values as a form of trim [17:22:47] as spurious as certain abort buttons [17:24:38] Once I fixed that, I flew a few landings in Orbirer2016 [17:25:17] the accleration issue was a pretty easy find too. AGS/PGNS disagreement by a factor of 2 [17:25:23] VI [17:26:28] and those N21 values I shared yesterday were likewise twice as big as they should've been [17:27:26] morning! [17:27:44] But I was finially able to answer the question I started yesterday with: "Can I fly a landing with no compensation, if I start PDI with a good state vector and allignment" [17:28:02] hey Mike [17:28:05] to which the answer is, "Yes" [17:28:09] Hi Mike [17:28:31] yeah I would have been surprised otherwise. If any real IMU would have been that bad it wouldn't have flown. Compensation or not. [17:30:15] That was my assumption too. [17:44:06] down to 2200 differences, with I think all major new features implemented. my Skylark is becoming pretty Skylark-y :) [17:44:17] ooooh nice :D [17:44:44] is that a lot of small differences or is anything going to still make a big difference instantly? [17:44:58] I think it is mostly a log section ordering problem [17:45:32] I have some evidence that e.g. log section A comes before log section B in Skylark, but other evidence that log section B comes before log section A [17:45:38] which means one of them probably split up somehow [17:45:57] ah yeah, not that surprising with Skylark [17:46:23] so at this point I think I need to catalog all of the log sections that are present in the three banks that still have major errors, and figure out how to get them in the right order [17:47:41] is it some rendezvous related sections? [17:48:55] the little bit I looked into it suggested that it was P20 and P34 [17:49:10] which got R27 and major changes, respectively [17:49:23] hmm yeah [17:50:09] I'm uploading the video of the docked Skylab burn. Not much fancy editing, just showing that the DAP modifications from the deorbit study work and docked SPS burns are stable. I'll link the document in the description of course. [17:50:19] nice :D [17:50:48] the burn is the shaping burn as that is what procedures were given for [17:50:58] there would have been a second burn for the actual deorbit [17:51:38] I'm not even really sure what the plan was. Quickly undock after the deorbit burn? Or a stabilizing SPS burn to stay in orbit [17:51:47] that's one of the reasons why they abandoned the plan, too risky [17:52:01] undocking has to work quickly or you are in trouble [17:54:01] hahaha yeah [17:54:44] but in either case, they better had that relative motion worked out [17:55:32] sadly this document is the only one so far I have found about this topic. And it doesn't give much more about the deorbit burn than the Delta V of it [17:57:17] personally I have a better feeling with staying in orbit. The residuals are a bit higher with the docked burn, so you might not reenter as precisely as usual [17:57:26] and you have to get away from Skylab, changing the orbit [17:58:03] if I had to design the sequence, I would have done undocking, bit of SM RCS to get away and then a quick SCS SPS burn for staying in orbit [17:58:35] although, the TVC DAP should be fine, only the CSM+"LM" padloads are changed I think [17:58:45] so CSM only TVC DAP would work [17:58:59] even without reverting the erasable changes [18:02:06] right, yeah that makes sense [18:03:23] sounds like quite a sequence of events :D [18:07:57] the thread on NASASpaceflight mentions someone has actually seen a document that gives a sequence of events... I'll have to look a bit for anything [18:08:04] about that [18:28:17] https://www.youtube.com/watch?v=_ay5bK091oA [18:32:03] that's a long title [18:33:09] :D [20:44:05] personally I'd be more interested in docked skylab burns in the other direction :) [20:45:51] right? [20:45:56] seems more useful [20:46:54] or maybe a module that gets delivered to Skylab to do the job [20:46:57] like [20:46:59] https://en.wikipedia.org/wiki/Teleoperator_Retrieval_System [20:47:39] but weren't there also talks to do it with Skylab 4 or an extra launch or something [20:49:24] ah but Skylab 4 actually did a reboost [20:49:27] just with the RCS [20:51:21] about 10 ft/s of reboost, 180 seconds burn time [21:03:41] night! [15:44:25] good morning [15:45:46] hello hello [15:46:31] I think I will send some NARA requests today [15:46:54] before that also stops being a source we can access, better not wait... [16:26:33] done. Somehow I have the feeling I am going to have to pay this time haha [16:27:38] 80 cents per page is steep, but what I requested is probably just under 100 pages, so I think I am prepared to spend that on this [16:28:49] yeah, that could be a lot worse [16:30:48] I would like it more if I was sure what I am getting [16:32:13] and often new documents make me find more new documents I want, so it's dangerous :D [16:38:46] morning! [16:38:56] very dangerous :D [16:39:01] what did you request? [16:39:30] hey Mike [16:39:43] not one I had put on the shopping list, but related [16:39:56] Basically the 1970-1972 version of the AEG [16:40:29] https://archive.org/details/68fm119images [16:40:41] this is the 1968 version that Ron scanned for me a few years ago [16:40:56] sadly it has a few too many errors to implement it [16:40:59] from just this [16:41:18] and the equations are super complicated and not really possible to derive myself without all the sources it would use, so [16:41:32] I hope from getting a revision of this that I can maybe implement it properly [16:41:49] https://archive.org/details/NARASWSelectedApolloBoxes/page/n851/mode/2up?view=theater [16:41:51] ah nice, yeah, hopefully they fixed the errors and didn't just add on to it haha [16:41:57] this is what I requested specifically [16:42:19] this version mainly added another atmosphere model for drag, but, it should include everything [16:43:05] sometimes it's also not really an error but a scan where e.g. a minus sign has vanished [16:43:27] did you request all three of these? [16:43:29] if I have 2+ copies of nearly the same thing I can surely extrapolate what I need :D [16:43:54] I requested 70-FM-166 including all change documents to it, knowing that there are at least 2 changes [16:44:14] cool [16:44:38] revisions are always good to have [16:44:51] Apollo 14 LOI Targeting? Too many issues, couldn't implement it [16:44:58] Apollo 14 LOI Targeting Revision 1? Flawless [16:45:55] they made that a whole new document with the revision [16:46:01] instead of change documents [16:46:17] shows the amount of things that changed in between :D [16:46:34] hehehe [16:46:51] the two change documents shown here look pretty extensive [16:47:27] yeah [16:47:31] almost would have been easier to just republish change 2 haha [16:50:16] With the DKI, they published a new document in late 1969. Then Change 1 and 2 are two small additions. Then Change 3 has the changes for the Skylab rendezvous profile! And there even is a 4th change :D [16:52:22] oh man, lots of revisions [16:53:05] they squeezed the Skylab profile into the DKI somehow. Probably not much cleaner than how I did it. [16:53:15] No wonder they chose to develop a more generic system afterwards [16:53:52] hehehe [16:54:31] The first Shuttle rendezvous profile they came up with in the early 1970s already couldn't be supported anymore. [16:54:49] although, I have been thinking if it wouldn't be possible somehow to trick Skylark into calculating it [16:55:25] I have some dreams of writing an EMP for it, but, that seems farfetched [16:56:22] hahaha oh man. using those building block functions you were talking about? [16:56:28] yeah [16:57:37] once your disassembly is done I will look at the code, just to get some ideas for something like that [16:57:43] well, speaking of Skylark, my build is now 100% correct [16:58:00] and I have built up questions to ask you haha [16:58:27] awesome [16:59:30] so I am actually able to keep the Artemis log section ordering, with only additions and removals, if I split up two things [17:00:42] first, the routine DELRSPL moved from after P20-P25 to before it. in Comanche and Artemis it's located in section P30-P31, and claims some relevance to those, but I haven't figured out how haha [17:00:51] it seems like it's part of R30, that only is used for P11 [17:01:16] so I'm wondering if they might have moved it into the R30 log section [17:03:46] ah, DELRSPL is what calls AUGEKUGEL [17:04:33] the wrapper function that calculates estimated splashdown coordinates [17:04:37] mostly for Mode III aborts [17:06:24] has DELRSPL itself been changed in Skylark or only moved? [17:07:43] only moved [17:08:05] yeah in P11 you would use it for a Mode III abort. Burn the SPS until splashdown range error is zero [17:08:27] splashdown coordinates are padloaded to the end of the East Atlantic abort zone [17:10:08] relevance of DELRSPL to P30-P31... [17:10:11] uhh [17:10:46] none unless the Apollo 1 software calculates estimates splashdown coordinates with it in P30 :D [17:10:55] estimated* [17:11:06] the problem could also be my P20-P25 log section. but DELRSPL needs to come before a bunch off stuff in P20-P25, and P30 itself needs to come after. so if P20-P25 has not been split up, then DELRSPL needs to be separated from P30 [17:11:46] haha yeah, and since DELRSPL has nothing to do with P30 as far as I can tell, I feel like moving it makes sense [17:11:52] R30 is V82, right? I would say DELRSPL belongs to R30 or P11 [17:13:00] yeah, I couldn't make a decision there. I went with R30, since it's only called from there [17:14:38] next, related question: I also am pretty sure log section P34-P35, P74-P75 split up. this is perhaps less surprising haha [17:15:46] specifically, R36 (stared by V90) appears to be in the same place as Artemis, while a lot of the code for the rendezvous programs themselves moved to later in the listing [17:16:20] so the question is, does R36 deserve its own small log section like R30 and R31, or do you think it would have been co-located with one of the specific rendezvous programs? [17:16:24] I can confirm that in the Apollo 1 software P32 (prethrust for deorbit) a noun is displayed with splashdown error [17:16:36] oh interesting! [17:16:46] so the version of this routine would be called there [17:19:46] is the Norton manual of any help with the code order? [17:20:03] zero help [17:20:11] Norton makes his own groupings and order :D [17:20:21] thematically R36 almost belongs to P38 [17:20:25] both for plane change stuff [17:21:48] ah, I have a related question for P38. in Artemis, P36 is located in the P30,P31 log section along with P30 and P31. I left it there in Skylark. so maybe either R36 should go into there, or P38 should come out and join with R36? [17:22:15] (I mean, left it there and renamed to P38) [17:23:29] what do you have in "P30,P31". P30 and P38? [17:24:15] currently, yeah. and I even renamed the log section to "P30,P38", to deconflict it with my new "P31-P33" log section [17:25:00] if it works out with the order of things, P30, P38 and R36 could be together [17:25:28] I think putting R36 in that section would probably solve my order problem. let me see [17:25:45] I mean, you would only ever call in all P3X programs except for P30 and P38 :D [17:25:50] call R36* [17:26:59] oh no, that doesn't work [17:27:14] R36 also needs to be before P20-P25 [17:27:25] and P30-P31 (now P30,P38) comes after [17:28:08] R36 is an extended verb, V90 [17:28:22] maybe somewhere where extended verbs are living? [17:28:50] oh that's an interesting point. extended verbs come very early in the listing [17:30:08] nope, that is too early :D [17:31:56] thematically R36 also belongs to V83, V85, V89 [17:32:06] which are routines R31, R34, R63 [17:32:53] in the GSOP section 5 R31, R34, R63 and R36 are in one category [17:33:53] and in code, R34 is in the R31 section, R63 is in P20-P25, and R36 is in P34-P35,P74-P75 [17:33:59] I would not have guessed relatedness there haha [17:34:41] yeah R31 and R34 are very closely related [17:34:55] just have a different register 3 [17:35:58] I might just leave it how it is for now, giving R36 its own log section where P34-P35,P74-P75 used to be [17:37:19] so then last question: I currently have the rendezvous programs divided in to sections as follows: P31-P33, P34, P35-P37 [17:37:40] does that division make sense? I'm pretty sure P31-P33 as a unit is correct based on code structure. P37 is probably the easiest to move [17:42:19] P33 is quite different than P31/P32, but I guess there are some similarities in the subroutines it is using [17:42:32] P34 alone makes sense [17:42:38] P35 and P36 are closely related [17:42:57] P37 is a bit of a transitional program, not doing that much. I think that can stay with P35 and P36 [17:45:23] great, thanks! [17:45:44] one thing they introduced in Artemis is that the CSI and CDH programs sort of do a V90 calculation automatically and apply the out-of-plane DV. They retained that for Skylark's NSR (P34) calculation [17:45:58] so I guess R36 (V90) and P34 also have something in common [17:46:12] hmmm interesting [17:46:42] even if that relation make sense, doesn't necessarily mean that this translate to code structuring :D [17:46:46] haha yeah [17:47:02] well, like naming things, without a source for log section names the best we can do is get close [17:47:15] log section names and division of the code into them, that is [17:47:45] that's my current source build, complete with HTML files, if you want to start poking around [17:48:20] I left myself over 200 FIXMEs related to naming things, organization, and comments that need updating, so there will be a decent bit of churn still before I send it in to Ron :) [17:49:01] oh yeah I will look at this, thanks! [17:50:27] ## FIXME REVISIT [17:50:39] what sort of thing has to be fixed there? [17:51:39] that's on COE right? I don't remember, past me should have been more descriptive. I think it might have been that Norton was using names instead of numbers for pushdown list locations, and I wanted to see if it made sense to use names in the actual code [17:53:05] yeah, COE [17:57:38] yeah, if you look at COE in Norton (pages REND-24 and REND-25), he's using a bunch of names like PDRM, PDRA, PDCA, PDCAD, etc. for pushdown locations [17:58:01] which raises two questions: 1. did the listing also use names? 2. if it did, do they have the PD prefix or was that a Norton-ism? [17:58:37] neither of which was immediately pressing for disassembly or source reconstruction, thus the note to come back and try to answer those questions :) [17:59:41] I see [18:13:35] RM is already the name of a variable, so I guess that one would need the PD prefix [18:15:47] see but one of them is called PD8D [18:16:00] surely they wouldn't have equated a variable called PD8D to 8D [18:16:22] so was that just Norton not being creative for a name, or what [18:34:56] yeah PD8D is a bit of a useless variable name [18:39:33] another weird example, PDA in REVUP [18:39:56] what Norton calls PDA never has a chance to even have a name, because it's never referred to by address. it's just pushed onto and popped off of the stack [18:40:02] so surely that one came out of Norton's head only [18:40:38] see why I was confused and just put the pushdown list address numbers on my first pass? haha [18:41:02] Norton is simultaneously incredibly useful and incredibly infuriating as a source [18:51:45] Norton knows better, and half the time that is actually true :D [18:51:54] hehehe [19:16:25] hmm [19:16:35] "it will cost $116.00 as there are 145 pages" [19:17:13] slightly more than I had hoped [19:19:36] also, I'm not sure these payment options will work for me [23:04:19] oooo Fortran :) [15:43:23] hey [15:43:27] thewonderidiot, I got the document! [15:49:38] hello [15:51:33] hey Matt [15:56:00] looks useful [15:59:20] nice! [15:59:30] that was a fast scan [16:00:36] I got the email 3 hours ago already haha [16:00:56] wow [16:01:22] I wonder if they just scan it right away and just wait for the payment... [16:02:34] oh interesting, maybe [16:03:46] is there code? [16:04:35] no, it's handwritten flowcharts [16:04:45] a lot better than the IBM RTCC flowcharts though [16:05:02] ahh [16:05:29] you said "Oooh Fortran" yesterday. The IBM RTCC flowcharts are badly handwritten, probably with the Fortran code as a source. Which is only a bad thing as variable names are all over the place :D [16:05:42] I had an interesting conversation with the author of the IBM 7094 emulator last night [16:06:04] aparently this person: https://github.com/VikParuchuri/surya [16:06:19] has taken an interest in this document: http://mwhume.space/Files/TranscribeMe/todo/19630001662.pdf [16:06:48] or more broadly, developing OCR tools to transcribe listings [16:07:53] might even work with that specific document [16:09:32] To run it, and I belive several other programs like it, we'd need to be able to generate the ephemerides tape in the format it's looking for [16:10:03] oh right [16:10:13] definitely a topic I have put some thought into already [16:11:07] as you can imagine, it's sort of hard to find a scan of a tape! [16:11:15] I haven't found any [16:11:20] just misc data for it [16:11:31] units used, conversion factors required etc. [16:12:02] I have written myself some "ephemeris tape" text files. With the correct data, at least for Orbiter [16:12:22] I believe the Orbiter time keeping is sort of similar to Ephemeris Time, so it's not that bad [16:21:05] I think in that n-body document, there is a good description of the format. for some of the others I think I'd have to probably reverse engineer the fortran read statements. Although I think this format is kind of a general standard [16:34:16] Do we currently load any ephemeris data for earth, moon, etc, or do we just use Orbiter data directly? [16:35:24] RTCC loads Orbiter data into tables at scenario load or when you change the launch day [16:36:11] the old version of the coasting integrator still calls clbkEphemeris for the Moon [16:36:20] so gets the data directly during integration [16:36:42] it also used to do that with the Sun, but now it uses a scheme similar to the AGC instead. [16:36:50] clbkEphemeris with the Sun was suuuper slow [16:37:08] calculations in the RTCC MFD took noticable longer [16:37:43] and those data tables are then interpolated by the RTCC [16:37:53] to get vectors at specific times [16:38:28] for the ATDP I saved some ephemeris data text files, so that it doesn't depend on Orbiter ephemerides [16:38:38] covering 1950 to 2000 [16:39:24] ahh okay [16:40:14] padload generator takes the Orbiter ephemeris files and does similar code to get data from it. So a different approach there. Doesn't depend on Orbiter API, but on Orbiter files. [17:32:43] morning! [17:35:23] indy91, excellent :D [17:35:49] I'm seeing lots of error corrections in the change pages too, so hopefully this one is less full of typos [17:46:32] also n7275, I think the software change for the PIPA bias scale factor is very trivial, so we could probably pretty easily make a slightly modified Comanche 55 for your Apollo 12 testing [17:51:57] Comanche 55 3/4 [17:52:25] thewonderidiot, yes and it's quite close to the 1968 version, so that's basically another version to check against [17:52:40] so hopefully I can get it working [17:52:55] awesome :D [17:53:43] ... but it will take a few days [17:54:20] not sure how fast Paypal is, but that is also done as of a few hours ago [17:54:43] I saw it, thank you! [17:54:55] no thank you :D [17:55:26] it's a bit weird, I think this document has Change 1 twice [17:56:01] hmm [17:56:18] did the base document already have change 1 applied, and then you got the change 1 pages separately? [17:57:22] nope [17:57:37] I'm pretty sure the order is: base document, change 1, change 2, change 1 again [17:57:39] the last 25 pages [17:57:51] with a slightly different cover or something? [17:57:53] oh weird [17:57:58] but it seems to be the same [17:58:03] I'm not going to convert that in lost dollars :D [17:58:08] hahaha [17:58:30] not a scanning error of course [17:58:37] that's how the document was bundled [17:58:48] also how Ron had scanned it [17:58:56] 3 change title pages [18:00:19] who knows, maybe having two scans of something saves me somewhere with a missing minus sign or something [18:02:23] good point :D [18:05:45] the first sentence of the document almost had me scared [18:06:27] "The AEG portion of flow charts 1 (ref. 1), 2, and 3 has been adequately documented in references 2 to 5 and will not be reproduced in this paper other than to note that.." [18:07:44] but they just mean there is no long discussion of what the equations do [18:07:59] not that there is in the 1968 version... [18:08:10] eventually I still want the Gemini version :D [18:08:55] that was first version from 1964 [18:09:44] oh boy [18:10:19] I don't even think all that much changed, not in the basic equations [18:10:31] the main developments in the 1968 and 1970 versions is for drag [19:34:42] hmm indy91, just double checking, but you already have the documents Ron has been adding to the document library the past couple of days, right? [19:34:51] there are some RTCC things [19:35:09] probably [19:35:17] haven't checked in a few days [19:35:29] there can always be something that has slipped through [19:38:36] yep, I knew about those [19:40:23] lol this is a first. somebody re-submitted to us a scan that I did of a document that I own :P [19:42:26] not that surprising at this point, the library is huge :D [19:43:34] a few Shuttle documents are new to me [19:43:42] hmm [19:43:48] apparently some documents come from this? [19:43:49] https://specialcollections.wichita.edu/collections/ms/87-08/87-8-a.html [19:44:08] really?? there's some good things in that collection that we don't have [19:44:47] https://www.ibiblio.org/apollo/Shuttle/IBM-82-SS-4556%20-%20Rev%204%20-%20Space%20Shuttle%20Orbiter%20Avionics%20Software%20-%20Programming%20Standards%20Document%20-%20Shuttle%20Flight%20Software.pdf [19:45:22] every page is marked haha [19:45:52] but yeah, there is stuff like "RTCC Computer Requirements for Project Apollo" [21:18:22] night! [21:23:22] thewonderidiot, that collection is where the ATMDC document is, that I've been emailing about [23:19:32] n7275: oh right! and they're not giving you too much trouble as far as digitization goes? [02:45:27] no, they've been super responsive. there's an ongoing construction project at WSU, and they said they should be able to get to it in early March [05:59:51] that is excellent news. I might have to make some requests there after you get the ATM doc :) [15:27:58] good morning [15:31:15] hello [15:32:29] I had a minor issue with the RTCC last night [15:33:56] oh [15:34:48] I wanted to compare state-vectors after launch for Apollo 12, but every time I tried, the CMC vector has a bunch of NaNs, which caused a crash [15:35:31] hmm [15:37:55] but you were already in P00? [15:38:36] the RTCC doesn't really pull the correct state vector that would be on telemetry [15:38:57] it uses one that is a byproduct of integration, because the EMEM address is stable [15:39:32] also, you did do TLM first and then move the CMC telemetry vector to the evaluation table, right? [15:39:49] I wonder if that second step if possible, even if no valid telemetry vector is available [15:42:00] hmm yeah there might be no error checking for this [15:42:16] So maybe, for whatever reason, you didn't have a valid state vector in the telemetry table [15:42:34] so the EV button moved a garbage or zero vector [15:46:45] I was definitely in P00, and yes TLM -> EV -> UV [15:47:17] just one more reason I ned to get telemetry working :) [15:49:31] weird [15:49:38] was this shortly after insertion? [15:50:01] do you have a scenario saved with that CCHE vector? It gets saved in the scenario [15:50:28] it was just before the first P52, so around T+40m [15:50:30] I wonder if, when P00 hasn't done an integration yet, that the vector would be bad [15:50:31] hmm [15:50:35] that's later [15:50:45] there definitely would have been an integration by then [15:50:46] and no, thanks to Orbitersound I do not have a saved scenerio... [15:51:14] but I will run it again tonight and try to replicate [15:55:18] in theory it could also be a RTCC integration issue [15:55:41] the state vector is correct when stored, but has gone bad when integrated to the comparison time [16:01:00] and I have definitely compared DC, IU and CMC in early Earth orbit many times before [16:05:04] CMC is always the worst [16:05:07] but not NaN bad :D [16:07:28] the crash can happen if the displayed number exceeds the MFD limits, I believe [16:07:43] I guess I need more protection against that... [17:36:47] morning! [17:48:28] hey Mike [17:49:06] I hear you have been asked to perform some very simple and basic AGC software changes for flying missions in 1974 and beyond :D [17:52:14] lol yup [17:54:35] the quickest way to success is if I come up with the calculations for changed constants. Exactly what I need, more projects for other people :D [17:55:34] but thinking about it, what would be the most convenient form of output of such data [17:55:45] to copy & paste into AGC code [17:56:33] COSI 2DEC* 9.996417320 E-1 B-1* [17:56:37] maybe like this? [17:56:40] I mean you could pretty much generate AGC code directly, yeah [17:56:41] a text file that has a bunch of these? [17:56:44] yup [17:57:29] but yeah I'm in the same boat lol. falling behind on projects, definitely not enough time to take on massive projects requested by other people haha [17:58:30] maybe it could be something added in the padload generator, because that already has a lot of the required functions for ephemerides, Earth/Moon rotation in Orbiter etc. [18:03:21] easiest GUI ever [18:03:26] all you input is a year [18:03:57] at least it is a project that can be done in several pieces [18:08:18] best part about the padload generator solution, what it calculates for Skylark Sun ephemeris padload is the exact same as for Luminary controlled constants [18:20:57] oh yeah :D [17:12:37] morning! [17:41:16] good morning [17:41:20] what's up? [17:48:07] not too much, taking a short break from work to check IRC etc. [17:48:52] oooo, I have a question for you [17:49:16] you do stuff with "punch cards" for pYUL [17:49:43] have you ever wanted for any kind of "metadata" for each card? [17:50:16] hmmm I don't think so -- what kind of metadata are you thinking of? [17:52:17] for some of the IBM 700/7000 stuff, it would be very useful to have mixed decks of binary and EBCDIC/hollerith/ASCII cards, but to do that I'd need some way of representing each card's format, and telling a parser what format it was [17:53:13] ohhhh weird [17:53:42] like for 704/9/90/94 cards, the binary cards are just 24x 36 bit words punched into the cards, 2 per line. that makes for complete ascii garbage [17:54:27] and doesn't play well in a mixed deck (mostly the line feeds, carriage returns, and form feeds) [17:55:54] mixed deck is crazy. I'm guessing the system reading them knows from the context which format is expected? [17:56:49] my gut reaction would be to put that info in column 81+, and strip it off when reading, assuming you're using regular 80-column cards [17:57:47] I really probably should store my cards in EBCDIC or something, but right now they're unicode :P [18:17:52] good evening [18:26:28] hi Niklas [18:27:12] good morning! [18:28:09] thewonderidiot, yeah the mixed decks I'm talking about would have multiple jobs, some of which would be compiling source, other would have input programs with control cards to run [18:28:37] I like the 81+ column idea. I'm going to work sumething up for that [18:33:13] huh, very interesting haha [18:33:38] indy91, how goes AEG? [18:34:05] I've encountered trouble [18:34:33] good news is, I found one bug in our current implementation which should improve accuracy by a few meters per orbit [18:34:56] but yeah, I found some differences between the 1968 and 1970 versions that don't make sense at all in the 1970 version [18:35:11] ... I kind of want the 1964 version now :D [18:35:19] hahaha uh oh [18:36:18] I also feel like there might be an error in the flowcharts [18:36:43] there are different routes through the equations for low vs. high eccentricity [18:37:13] and it seems like, different than in the 1968 version, before it splits off the 1970 version already does some low eccentricity calculations [18:37:39] almost seems like duplicated code, but not quite [18:37:53] so I don't think I will be able to use this 1:1 [18:38:25] but I still have some way to go [18:38:32] so maybe I can find the issues [18:54:41] it's probably a false hope that the 1964 version is any less error free haha [18:54:56] but it might give derivations of what it's actually doing [19:00:02] does NARA have the 1964 version? [19:00:12] likely [19:00:23] I had Ron scan some documents from 1964 already [19:00:51] and I know the name and memo number [19:01:14] always a chance they are missing one in the series of memos, but, probably a good chance they have it [19:02:18] they had 64-FM-55, 64-FM-57 and 64-FM-74. So why not 64-FM-50, too? :D [19:02:43] oh yeah, seems likely haha [19:03:48] something tells me it isn't 20 pages, like those 3 were on average [19:04:04] but probably not 150 pages, unless there are lots of change documents [19:04:29] well I'll try with these later versions for now [19:04:37] surely I can make it work with that [21:52:48] night! [18:14:16] morning! [18:25:19] good evening [18:25:52] hey [18:27:58] what's up? [18:31:13] slowly chipping away on Skylark cleanup [18:31:23] lots of nouns and erasables to update comments for [18:31:51] and you? [18:34:26] busy day, but now I am going to put a few more things in place to calculate the yearly rope constants, in the padload generator [18:34:48] already did a good cleanup of the padload generator for it. Loading those constants (for known ropes) from a file [18:35:05] because after generating them the padload generator still needs to support it for padloads [18:35:45] so you would have to add a data set of constants so the padload generator can generate padloads with it [18:36:26] and after that... I think what I will try with the AEG is to use the 1968 version as a baseline, with known fixes from the 1970 applied to it. Maybe that works better. [18:36:52] at least I fully trust the flow of the flowcharts in the 1968 version :D [18:37:16] what's much better in the 1970 version is variable names [18:37:44] where the 1968 version has e.g. a'', with the '' barely visible the 1970 version would have ADP, double primed, written like that, ADP. [18:38:22] yes that sounds much better :D [18:38:24] so there is no ambiguity or printing/scanning issue that can make something like the double prime disappear [18:39:35] so maybe the best of both worlds is the solution [18:41:32] it is an interesting problem, trying to make something that works by combining two sources that are both wrong :D [18:46:34] but hopefully wrong in different places :D [18:47:21] I mean, this whole issue is why analytical propagators have fallen out of favour. Maybe except for the SGPs. The equations are too complicated to understand and you can just throw a numerical solution at the problem, with the computing power today. [18:47:46] But there are definitely some desirable properties beyond that, which is why I like the analytical solutions [18:47:58] it gives mean orbital elements, for one thing [19:04:51] hey guys [19:06:37] hey Matt [19:08:21] hey [19:17:33] thewonderidiot, I had never noticed it before, but the Luminary 1E GSOP section 5 is missing a page [19:18:14] oh weird [19:18:31] which page? [19:18:58] 5.5-24 [19:19:53] oh well, I don't think it changed from previous versions [19:20:46] hmmm, annoying. but good that it's something that probably didn't change [19:24:12] yep [20:06:24] I'm making some slow progress on the telemetry processor [20:06:33] thewonderidiot, they scaled something in the LGC solar/lunar ephemerides to units of days, so that they could put 1/365 and 1/27 in the controlled constants, because that's how many days the Sun (well, Earth) and Moon take for one orbit [20:06:49] kind of reminded me of some Block I stuff :D [20:07:12] n7275, oh nice! What are working on right now? [20:07:18] are you* [20:10:22] I have a function that looks like the winsock send/receive functions. [20:11:15] my plan is for the CSM and LM to call these functions, which will fill up a buffer [20:12:03] then a telemetry processor thread will read data out of the buffer and sync/sort [20:13:13] I have it living in the MCCVessel right now, but that's just a convenient home for it right now [20:13:49] indy91, that is reminiscent of our adventures in Sunrise/Corona orbital integration :D [20:32:24] n7275, hmm, so should we then change the telemetry clients to use this system as well? [20:32:33] or is your plan to replace the telemetry clients [20:33:16] I sort of like the system with TCP/IP in that NASSP code just needs to send out the data and then can forget about it. No buffer in NASSP/Orbiter needed. [20:34:15] or is this supposed to, in the long term, be the in-sim, MSFN processing stuff [20:38:29] I want to still support the telemetry clients for the time being. depending on feasibility/performance I'll either leave toe tcp send/receive in place or have the telemetey processor "forward" telemetry and commands [20:39:31] *the [20:41:25] I'm not hard set on any particular way though. [20:48:42] yeah, difficult to say what's best. It would be nice to access telemetry data with the RTCC MFD eventually. For display and processing of data. But in a way it almost seems better to have this in a RTCC that doesn't even run within Orbiter... [20:54:26] there's a bunch of time/data synchronization needed for that though. I wonder if a "RTCC client" might be better though. [20:56:44] sending in MEDs over a telnet collection would be pretty easy [20:58:50] no that wouldn't be very realistic. MEDs are going from a RTCC console to the RTCC computer. There should be no sort of client there, it needs to run together. [20:59:01] any interface should be between "space" and "MSFN" [21:02:11] RTCC being part of MSFN in that sense [21:02:37] true, the consoles are about as close as physically possible to the computer [21:02:39] https://lh6.googleusercontent.com/proxy/nLMARSl01e_IVWV8cvL0FTAUGmJH2GFCIOn32Vrb1DYvB8neJ7jZT8byhaEq9OI2E9wr2PcEdDAV0YCFHPc72ErOWceAyAaqqwCcDBGaOV18PQ [21:04:32] of course a RTCC completely external to Orbiter would basically mean writing it from scratch, without the Orbiter API [21:04:37] so how do we make something that allows an "external RTCC" without "needing" it [21:04:41] so that might not happen :D [21:05:24] so your approach is probably right, do some processing in the MCCVessel [21:18:49] night! [16:23:46] hello [16:35:15] hey Matt [16:39:23] as is typical for me, I've got myself stuck right at the start of a project [16:40:38] I had to write a script just to understand one single equation in the AEG drag calculations. Does that sound any better? :D [16:40:51] oh and I am not understanding it yet haha [16:43:55] sounds like we're in similar boats haha [16:45:26] I can make pretty graphs already at least [16:47:15] morning! [16:47:40] https://i.imgur.com/kJciVV9.png [16:47:52] semi-major axis decay, AEG takes no drag into account [16:48:30] oooo very cool [16:49:26] https://i.imgur.com/d4bCbXN.png [16:49:38] very simplified equations to take drag into account [16:49:51] hey Mike [16:53:27] Apollo 7 post SPS-7 orbit, 90 x 230. With exaggerated area for more visible drag effect. [16:54:03] this is as far as I understand every equation, so now I am trying to figure out how they calculated the altitude for the density :D [16:59:25] nice! [17:13:54] I'm getting a pointer to the MCC vessel, but for some reason it appears to be some strange uninitialized shadow copy [17:15:03] in which class are you getting the pointer? [17:15:15] PCM [17:16:34] I do something similar in the antenna classes, and have no issue there. [17:17:10] oapiGetVesselByName("MCC") [17:17:11] ? [17:17:45] which PCM function. Could it be that the MCC hasn't been created when you call this function? [17:18:03] in the Saturn class the MCC pointer is gotten in clbkPostCreation for that reason [17:19:33] and the antennas get a pointer every timestep, in case the MCC gets deleted I guess :D [17:19:48] I'm doing the same thing. [17:20:07] if(!mccv) {get pointer} [17:20:32] else { send data} [17:20:52] and mccv is a class variable and not local? [17:22:05] yes it's a member of PCM [17:22:07] https://github.com/n7275/NASSP/blob/627fc867ec017be54451d030292e26e1b6d50c79/Orbitersdk/samples/ProjectApollo/src_csm/csm_telecom.cpp#L4963 [17:23:22] i have a debug string in that TelemetryDownlink funtion, and it can see data coming in [17:24:24] and I can do MCCV->CSM_TelemetryBuffer.push_back(value) from the PCM class. [17:25:44] and it actually puts data in the buffer....but TelemetryDownlink throws an access violation exception when it tries to access members in MCC [17:26:42] is MCCV initialized? [17:26:50] to NULL I mean [17:27:03] in the PCM constructor [17:27:43] ah yes, I see it now [17:27:51] aaaah wait [17:27:54] MCC class [17:27:56] not the same as [17:27:58] MCCVessel class [17:28:08] any confusion there maybe? [17:29:20] ahh, that is very likely the issue [17:30:11] I am confused though. IsVessel checks on [17:30:21] _stricmp(classname, "ProjectApollo/MCC") [17:30:30] is that the vessel or not... [17:32:43] maybe that is actually the name of the vessel config file [17:32:52] which is under ProjectApollo, MCC.cfg [17:32:59] but says [17:33:04] ClassName = MCCVessel [17:33:23] we are at least typically using this function to check on MCCVessel [17:33:49] not sure if you want to access MCC or MCCVessel but [17:33:51] MCC *MCCV; [17:33:57] would have to be changed if you want the vessel [17:34:11] if (utils::IsVessel(pVessel, utils::MCC)) [17:34:12] { [17:34:13] MCCVessel *pMCCVessel = static_cast(pVessel); [17:34:14] if (pMCCVessel->mcc) [17:34:15] { [17:34:17] pMCC = pMCCVessel->mcc; [17:34:20] } [17:34:21] } [17:34:22] is how the Saturn class does it [17:34:24] last step gets the MCC point from the vessel [17:34:29] pointer* [17:35:46] yeah that should do it [17:36:39] I just didn't realize those were seperate classes [17:37:18] that works! [17:37:43] MCCV = ((MCCVessel*)pVessel)->mcc; [17:38:10] now I just need to rename MCCV to be something less wrong [17:38:51] and I suppose do something with all this data I'm pushing into this buffer [17:48:29] thank you [17:48:43] sure! [18:13:46] so indy91, does the AEG drag calculation take into account orientation? [18:13:57] nope [18:14:03] CD = 2.0 [18:14:07] buuuuut [18:14:29] the 1970 document has something about Skylab [18:15:11] uhh, the document isn't OCRed very well [18:15:36] https://i.imgur.com/krPRYwG.png [18:16:02] but other than this, the AEG (including during the Shuttle program) had a fixed CD, fixed drag area, fixed mass [18:16:19] the Shutle FDO Handbook has a procedure to iterate on downrange because of this [18:16:50] calculate the OMP solution, which uses the AEG. Move the burns to the ephemeris, which has venting tables, attitude timelines (with attitude dependent drag) [18:17:10] so you get a difference between the targeting and actual results [18:17:22] So they needed to bias some OMP inputs because of this... [18:17:33] difference between AEG and numerical integration basically [18:17:44] ahh, cool [21:41:17] night! [15:34:27] fun: one of the documents Ron recently uploaded is a "MCC Trajectory and Orbit Determination Software Workbook" from 1991, which has pages about the AEG [15:34:41] not so fun: The document is missing those pages... [15:34:53] The PDF* [15:35:27] it's funny reading the document, it has system constants and function names that are still identical or nearly the same as in the RTCC in 1968 :D [16:51:56] thewonderidiot, do you know who is sending a lot of these documents to Ron lately? Isn't that person on Discord, too? Is it Voteins? [17:05:01] morning! [17:05:10] I'm not entirely sure, but it might be, yeah [17:12:26] ah ok, I have a question about it haha [17:27:05] if "MCC Trajectory and Orbit Determination Software Workbook" came from NTRS I would try to use the NTRS ID to search for similar documents [17:27:25] but maybe they would be in the queue to be uploaded anyway [17:50:44] hello [17:54:09] hey [17:56:49] hey Matt [18:23:01] I would not be surprised if they were running the exact same software in the early 1990s just under MVS and on IBM 3083s [18:26:27] indy91, I'm nearing the end of Skylark source reconstruction, and have just a few functions I need help naming. first, this function in P31-P32: https://github.com/thewonderidiot/virtualagc/blob/Skylark48/Skylark048/P31-P33.agc#L736 [18:26:30] I had found a document about that actually [18:26:40] Norton shows it called twice here, once with Rac and once with Vac: https://i.imgur.com/jMQL4i5.png [18:26:47] I think in the early 90s a lot of Fortran code got converted to Ada(?) [18:26:51] including the OMP [18:28:56] thewonderidiot, https://i.imgur.com/MTMT2pa.png [18:28:58] GSOP [18:29:41] this projects the CSM state vector into the plane of the Skylab state vector [18:29:52] so that all the P31 iterating is done conic and in-plane [18:30:55] "When phase match is successfully completed, the CM state vector is [18:30:57] rotated into the plane of the WS orbit." [18:31:23] so the name would be something with "project" or "rotate" or so [18:33:00] VECPRJCT or so [18:33:42] awesome thanks! [18:34:39] and then I have two functions in P35-P36 [18:35:07] these two: https://github.com/thewonderidiot/virtualagc/blob/Skylark48/Skylark048/P35-P37.agc#L889 [18:35:12] the second even calls the first :D [18:35:55] n7275, https://ntrs.nasa.gov/api/citations/19960022635/downloads/19960022635.pdf [18:37:12] the first one is constructing the matrix as shown here (but not going so far as to do the matrix multiplication): https://i.imgur.com/DEQo4ZJ.png [18:42:52] I'm a bit confused by this [18:43:02] it calculates DELVEET3 from DVLOS, not the other way around? [18:44:27] in this specific case at least, it seems so :D [18:44:35] this is P36, happening on pdf page 163 here: https://www.ibiblio.org/apollo/Documents/programmed-guidance-equations-skylark-001.pdf#page=163 [18:45:31] ooh [18:45:45] I was only looking in P35 [18:46:01] I think this is P36 only [18:46:05] I see it in the GSOP now [18:46:26] in P35 you could manually change the LVLH DV and it would process it [18:46:36] in P36 you change it in line-of-sight coordinates [18:47:11] I think that's a change from Artemis [18:49:24] this is how Norton shows the second function, with the first inlined: https://i.imgur.com/KKStoxM.png [18:50:45] that is in S34/35.3 on pdf page 139 here: https://www.ibiblio.org/apollo/Documents/programmed-guidance-equations-skylark-003.pdf#page=139 [18:53:31] so U35,3376 [18:53:32] just builds the matrix for the conversion from LOS to inertial [18:53:46] P36 own code then does the matrix multiplication [18:54:24] yep [18:54:31] and even if N59 wasn't changed (triggers the calculation in U35,3376), the second function converts inertial (DELVEET3) to LVLH and LOS [18:55:02] hmm [18:55:53] oh I see [18:56:01] MXV vs. VXM [18:56:16] both P36 and the second function use the first function to build the matrix [18:56:23] but then convert in different directions [18:56:33] P36 (if N59 was changed) converts LOS to inertial [18:56:44] second unnamed function converts inertial to LOS [18:58:10] so the first function definitely needs to be named something like "builds matrix" [18:58:50] the LOS matrix essentially [18:58:57] yeah. probably a name somewhat similar to LOMAT, which is located just above and also used by the second function [18:59:52] hmm, where is LOMAT in Norton? [19:01:05] in the same snippet as the last picture I linked -- also inlined without using its name [19:01:30] it's building the first of the two matrices used by U35,3411 [19:01:35] ah [19:01:44] that converts between inertial and LVLH [19:02:07] why did they name that one LOMAT haha [19:02:18] hahaha good question :D [19:03:11] how about LSMAT :D [19:03:40] that works! [19:04:55] in Artemis NOVRWRT does a lot what your second unnamed function does [19:05:58] but I can't decipher that function name enough to say that it should rather belong to the new function [19:06:17] second function calculates DVs in two coordinates [19:06:25] DVCONV or so [19:06:45] I can decipher it [19:07:04] S34/35.4 STQ SETPD # NO ASTRONAUT OVERWRITE [19:07:09] NORMEX [19:07:13] 0D [19:07:19] GOTO [19:07:23] NOVRWRT [19:07:27] "no overwrite" [19:07:28] so I don't think it belongs :D [19:07:29] yeah [19:08:34] I should kill that label entirely, it's not called and the line it was on changed [19:09:09] DVCONV sounds good for the function, though, thanks! [19:09:19] I bet they changed P36 in order to input chart DVs [19:09:28] that's it for now -- I might come back with more tomorrow [19:09:50] am I correct that P35 does not use U35,3376? [19:09:54] only P36? [19:09:58] directly I mean [19:10:02] not via the second function [19:10:49] yeah that's correct, P35 does not use it directly [19:11:03] ah good [19:11:07] then the GSOP is not wrong :D [19:11:20] I was looking at P35 at first and got confused about the order of conversion [19:11:25] direction rather [19:12:11] (SKYLARK) PCR-SL421, "Allow Astronaut Overwrite of N59 (ΔvLOS) in P36" (not dated) [19:12:33] that's the change [19:14:42] and that feature would have been used in some IMU failure cases [19:14:47] inputting chart solutions [19:15:38] however that would be useful... [19:15:48] maybe you could use the P36 calculated attitude with the GDC or so [19:15:56] oh speaking of, did you ever fly your IMU failure rendezvous? [19:16:04] not yet! [19:16:16] I definitely will though [19:16:18] should be quite fun [19:25:08] I'm looking forward to hearing about it :D [19:37:35] a complete Skylab rendezvous book would be nice, with the full IMU failure procedures [19:37:40] we only have that for ASTP [19:59:04] I will now do a test of this P52 GDC procedure though. With EMP :D [20:00:50] with our (not really drifting) GDC I bet we can get 0.1° precision there [20:22:23] the procedure and EMP worked nicely [20:22:35] it can't be quite as accurate as I thought [20:22:50] you manually load N20 angles using the GDC angles [20:23:09] and those N20, IMU attitude, is what the CMC uses for its attitude for the marks on the stars [20:23:18] but you are drifting in the SCS deadband [20:23:38] so it is off a bit from that drifting [20:23:57] had a 0.14° star error, a lot higher of course than during a usual P52 [20:24:07] procedure says to accept up to 0.4° though [20:56:38] cya! [15:30:09] good morning [15:37:53] hey [15:38:35] I made some more progress on my telemetry processor [15:38:58] still very proof-of-concept [15:40:22] but the processor thread has no trouble reading out of the incoming buffer and into a circular buffer [15:41:04] 40x in hbr and I was still getting 200fps [15:41:51] ah nice. Yeah I guess this can be done quite efficiently. [15:41:57] I guess I shouldn't be supprised that modern computers can move hundreds of kilobits/sec around in memory [16:01:13] I mean it can still snowball into a huge amount of data [16:01:36] like if we were to simulate the onboarddata tape recorders... [16:03:01] I found a nice JSC document number in the Shuttle FDO handbook. I give it about 0.0% to 0.1% that it could be FOIAed :D [16:03:20] I think it would be similar to the IBM RTCC flowncharts [16:03:31] JSC-118286, CCS Trajectory Operations Design Specification [16:04:48] yeah there is no chance, that screams "internal use only" [16:05:05] even if it's outdated, from the Shuttle days [16:18:24] Yesterday: [16:18:27] https://i.imgur.com/d4bCbXN.png [16:18:52] Today: [16:18:59] https://i.imgur.com/i7ee1uD.png [16:19:10] There is still something wrong with using the right altitude for density [16:19:18] both other than that I solved most issues with this [16:19:29] using almost exactly what the IBM RTCC document has [16:19:59] 16km position error after 10 hours, but this case is with an exaggerated drag area [16:21:10] 1.6km error with normal CSM area [16:21:19] and I bet it goes below 1km if I figure out this altitude [16:44:34] that's pretty good [16:45:16] I should add an ASTP test case [16:45:31] that's what this drag stuff will be especially useful for [17:16:49] https://i.imgur.com/rryu3Gz.png [17:16:55] 24h in an ASTP typical orbit [17:17:02] not great yet, but, could be worse [17:23:57] morning! [17:33:06] hey Mike [17:34:43] hi Mike [17:41:55] finish line for Skylark? [17:47:47] it's coming up very quickly :) [17:48:16] doing final comparisons with Artemis, and making sure I didn't miss any restart points or things like that [17:48:59] then I just need to proofread all of the massively changed comments (assembly & operation information, flag definitions, etc.) and flesh out the header blurbs :D [17:52:37] as long as the diff to the rope is zero in the end, that is the important part :D [17:53:07] haha yep. and I think I'm done changing actual code so that should be easy [17:53:30] I say that but then I made a bunch of changes to the W-matrix definition in erasable assignments last night [17:53:40] so we will see haha [22:13:18] night! [17:04:48] hey [17:06:19] morning! [17:06:34] good morning guys [17:06:42] err.. afternoon [17:07:02] good evening! [17:07:11] I am allowed to say that just after 6pm, right? :D [17:07:59] I'd say after 5pm, personally [17:08:25] getting dark outside, it definitely works for me [17:46:08] what's up? [17:52:48] well, Skylark is done. and I have spent almost all of my free time working on it since the New Mexico trip, with only one or two days "off". so I'm trying to decide if I want to pick up another project right away, or just chill out and recover for a while haha [17:55:39] taking breaks is good [17:55:43] ...I hear [17:55:47] hahaha [18:00:03] you can take over one of my projects if you need one :D [18:00:40] calculating the constants for other launch years for the AGC for example. [18:01:57] no thanks, you can have that one :D [18:05:41] guess what, I (for once) could use a MIT memo from NARA for that [18:06:22] https://www.ibiblio.org/apollo/NARA-SW/IndexOfNaraBox209G.html [18:06:28] one I saw here [18:07:00] oh, which one? [18:07:15] R-385, [18:07:23] Inertial Orientation of the Moon [18:07:26] ,Richard C. Hutchinson, [18:07:28] October 1962 [18:08:34] that is being referenced in the GSOP section 5 for the constants [18:08:39] among other things [18:08:55] I don't really know how useful this would actually be [18:09:10] might just be a derivation of the system they eventually put in the AGC [18:09:19] more than how to calculate numbers for it [18:11:59] right. kind of like the Compleat Sunrise's description of the orbital integration software [18:12:11] a nice derivation, but not actually useful at all for understanding the implementation :D [18:16:06] I understand the implementation, I just don't know yet how to convert Moon position and orientation data into the angles and rates of change of the angles they put in the constants [22:20:37] night! [17:37:24] morning! [17:37:57] hey Mike [17:38:02] what's up? [17:38:26] just looking at how the S-IVB PTCs on the last few missions were initiated [17:38:34] I think we don't need a lot to add that [17:39:29] ah nice [17:15:37] hi Nik [17:17:26] hello hello [17:18:20] 17 open PRs. I am trying to work against that this weekend :D [17:23:32] oh boy. want some more :) [17:24:02] no [17:24:03] I'm actually going to close that telemetry one since its for the wrong branch [17:24:08] yes :D [17:24:15] down to 16, well done [17:26:26] I'll try to do more IMU testing on Saturday. Hopefully we can build some confidence in that. [17:29:39] morning! [17:29:44] hi Mike [17:36:25] hey [17:36:59] n7275, I still have to implement the GUI for the padload generator for the bias compensation [17:37:25] and then add numbers for more missions [17:46:35] yes, we're still a way off from being ready to merge [19:56:05] indy91, the apogee PR should be good to go right? I was just reading through the code [19:56:33] yes, it is simply two small equations fixed. In both cases the original source I used for them was wrong... [19:56:59] I tested it all in MATLAB scripts, that's where the graphs come from [19:58:49] and the AEG thing, the error was about 1/4 of the J4 non-spherical gravity [19:59:00] I had trouble creating graphs for the small difference that makes :D [20:03:32] I had a sign error in my first implimentation of Pines, that someone eventually pointed out on github. from just looking at the code I would've assumed it would result in a huge difference, but was much more like: different, but not in a way anyone would notice or that I can really plot [20:06:14] that's especially bad with these AEG equations. They are so complicated and one sign difference can get you an error that is totally acceptable, but will be there forever unnoticed... [20:09:17] I screwed up not getting more documents from UHCL when we had the chance. One time I wanted to request the Flight Controllers Operations Handbook. But just to see what it actually contains I only requested a change document, which had a table of contents. [20:09:30] Didn't look terribly interesting, so I never requested the whole thing [20:09:53] but now I see something like "SOP 2.6 PIPA and IRIG bias correction" [20:10:04] probably procedures and guidelines for changing them during the mission [20:45:30] UHCL is not sharing anything anymore right? [20:47:01] I had requested a Skylab G&C Checklist [20:47:11] ... before I knew that ThatRedMelon already had a ASTP one :D [20:47:28] and yeah they said they are currently not simply sending out scans of documents [20:47:54] "Due to an ongoing matter, NASA has requested that the UHCL Archives not provide access to the JSC History Collection or fulfill digitization requests for documents at this time. This is an indefinite situation, and we cannot provide you with a date when the collection would be made available again to researchers. We apologize for the inconvenience, but this situation is out of our hands. We will let you know when access has been approved to open [20:47:57] back up for research to the JSC History Collection. As such, I can’t guarantee that we will be open to having onsite researchers by mid-March. We will also be requiring all researchers requesting documents or scans to complete a request form in the future for any request." [20:49:12] it kind of feels like we're headed into NTRS debacle 2.0.... I should refresh all of my backups this weekend [20:49:45] Some day someone might also not like what Ron is doing with the website. Deserves a backup, too. [20:49:56] yeah [20:50:11] I have an offline back of all of the virtualagc documents [20:50:36] and the stuff on Archive(.)org [20:51:01] maybe the UHCL thing is ITAR related. Or somebody thinks there is the possibility of something potentially is covered by ITAR, maybe. Which has of course (frustratingly) the exact same end result. [20:52:31] maybe I should just print everything [20:53:17] "Look at me, I am the JSC History Collection now" [20:56:56] I bet Mike would still have more physical documents than me though [20:57:11] :P [21:02:08] my first job in 2004 was working at a library, you know, if "The Stewart Foundation" ever needs to hire anyone :) [21:02:37] hehehe [21:05:45] I scanned documents all day for a few weeks between high school and university, at the place where my mother worked. [21:59:05] night! [16:23:49] morning! [16:40:25] good morning! [17:03:52] what's up? [17:43:06] oh not too much. Was trying to re-fly Apollo 11 with my IMU drift code, but I can't seem to get out of my own way this morning [17:43:13] er...afternoon now [17:51:55] hahaha [20:04:37] okay, finially launching [11:47:14] good morning [12:10:47] hey Matt [12:10:57] n7275, how did you get into a 100 x 150 orbit? https://github.com/orbiternassp/NASSP/pull/1030#issuecomment-1986985078 [12:30:25] that is a very good question [12:32:20] I have several qyicksaves I need to look at [12:36:54] I had hoped you did a SPS maneuver or something [12:37:49] maybe it's one of those cases where LVDC navigation or S-IVB venting does what it wants and I can never figure it out... [12:45:09] was this in OO? [12:48:36] nope, Beta [13:13:30] running rhrough a launch again [13:26:00] 103.8x102.2 [13:26:03] weird [13:27:18] You said you have some quicksaves? I'd like to check when the trajectory became wrong [13:27:25] immediately with orbital insertion or later [13:29:32] of course a buggy amount of venting thrust could cause this, and in the process also perturb the LVDC state vector when it's not running its accelerometers anymore. [13:29:34] oh it was immediately [13:30:05] so likely a LVDC navigation issue you think? [13:30:11] state vector being off? [13:30:26] some transient before orbital insertion [13:31:31] I was watching N44 after +10 minutes. It was like the SIVb just shutdown late [13:32:26] it kept on going so confidently, I first opened PAMFD to check, and then had a moment of doubt and looked at mission documentation [13:33:08] how soon after insertion do you have a save? [13:33:30] I find it unlikely that this is a guidance issue, although it's possible of course [13:33:47] more likely the LVDC state vector which has given me trouble [13:33:51] well never myself directly [13:33:58] because I never have any issues with it [13:34:08] just in scenarios of other people [13:34:35] lvlog from launch might have revealed something, but you probably didn't save it [13:36:10] ahh nope. that would've been good to remember. I thought it was a different problem at the time [13:36:21] looks like I have one at +17m [13:36:33] and -25s [13:37:27] I'm sure the T+17m one will have a terrible IU state vector [13:42:50] yep, LVDC thinks it is in a 100 NM circular [13:43:33] I have done a bit of debugging on scenarios with issues like this in the past [13:43:53] my conclusion was that the IU IMU attitude is fine [13:44:01] so it's not a misalignment [13:44:11] it's something with accelerometer processing [13:44:34] possibly something at liftoff already with touchdown point weirdness or so [13:46:19] is there any way my code is messing with the lvimu class? [13:49:26] I don't think so [13:49:31] you made no changes to the InertialData class [13:49:36] that could have affected it [13:50:55] maybe I need to run at 60 Hz more often [13:51:29] with the changes I had made to the LVDC class, make it more task based like the real thing (from Apollo 12 on), the LVDC timestep runs at a regular 1/100 second timestep [13:51:41] code similar to our Virtual AGC timestep [13:52:08] so I am wondering a bit if it makes a difference if the framerate is higher than the LVDC timestep rate or lower [13:52:19] I typically run 144 Hz so it is higher for me [13:52:59] but maybe there is a random bug that can only happen below 100 fps [13:53:29] I have a 60Hz monitor, but I leave vsync off [13:53:43] I'm typically around 90 in VC [13:55:32] that's another thing, maybe there is some bad code that uses the previous timestep length for something critical [13:55:47] then a non-regular frame rate could cause an issue [13:56:51] I know that the 100 Hz LVDC timestep can cause instabilities in measured acceleration [13:56:58] maybe it's related to that [13:57:14] but the real LVDC would have to deal with something like that as well [13:57:46] it all comes down to one thing, i need to make my performance bad on purpose I think :D [14:00:19] open like 3 tabs of chrome [14:01:52] haha [14:02:00] frantically scrolling around in Google Maps works pretty well, too [14:02:19] ... [14:02:25] guess how my SM looks [14:02:34] uh oh [14:02:43] good old missing panel [14:03:33] good to know we haven't accidentally fixed that [14:05:59] while I have that I might as well try to debug it a bit again [14:06:22] it's basically just that the alpha channel (transparency) is set to 0 [14:06:31] where the mesh file has 1.0 [14:06:48] I can even make the panel appear again with the in-sim debug controls [14:06:51] D3D9 [14:08:41] I guess potentially SetMeshVisibilityMode [14:09:17] maybe a case where that is called with an uninitialized mesh index [14:10:09] I have to believe it's related to the bug where if we don't close down Orbiter completely, all the animations get messed up...sometimes [14:12:34] oh wow, we call SetMeshVisibilityMode a lot [14:17:21] what mesh is that? [14:30:43] pretty sure it's SM-Panel4.msh [14:33:32] it also happens to be the Apollo 13 panel. but I think we proved out that it wasn't that code. [14:38:28] oh I have the advanced DX9 debugging option enabled [14:38:39] meaning I get a 13MB file when I close the scenario :D [14:38:44] maybe something is in there [14:40:34] when I try to relaunch a scenerio without reloading Orbiter I get an access violation from d3d9 client [14:41:30] yeah [14:41:42] I am sure we do a bunch of bad things in code related to that [14:42:26] at location 0x024, so it passes if(!nulptr) [14:44:26] is that Orbiter or NASSP code? [14:45:04] no way to tell, but almost certianly NASSP [14:45:32] but what sort of access violation is it, can we tell? [14:45:38] something not being deleted [14:45:42] ? [14:45:52] I know that the Orbiter log also says something like that [14:47:01] I don't think we can tell. just that D3D9 client got passed a bad pointer [14:48:07] regular Orbiter scenerios I can go back simulation -> launchpad -> simulation many times; no issue [14:48:13] yep [14:48:41] 000004.627: D3D9: ERROR: [Failed to Reset DirectX Device] (Likely blocked by undeleted resources) [14:49:54] 000000.020: D3D9: ERROR: D3D9ClientSurface: GetDC() Failed [14:50:27] debugging says it's related to the EMS scroll, but I am not sure I can trust this debug location while in release mode [14:53:04] 000000.020: D3D9: ERROR: Surface name is clbkCreateSurface Handle=0x23DCFFD0 (2500,145) [14:53:08] definitely the EMS scroll [14:53:49] have you loaded Apollo 5 recently? [14:54:07] nope [14:54:10] oh never mind, I was paused [14:54:54] the LM starts out on the ground for the first timestep. its back where it belongs now. I'm sure thats fine [14:59:10] hmm I guess so [14:59:18] I'll look a bit for undeleted resources [15:04:44] hmm, I got it to happen in > orbiter.exe!004e3a1b() [15:05:07] i'm not sure that's significant [15:10:03] for me it was at the oapiGetSketchpad for the EMS scroll [15:11:54] aside from the size nothing too unusual is done with the EMS scroll [15:12:15] the weird special code for it is all for writing a bitmap to file, on request with the PAMFD [15:43:18] oh I might have found something [15:43:37] fonts, brushes, pens are released in ExitModule [15:43:43] not in the Saturn destructor [15:44:04] maybe ExitModule is only called when you close Orbiter? [15:45:14] ah, right now you couldn't/shouldn't delete them in the Saturn constructor because they are global [15:45:17] not part of the Saturn class [15:47:36] confirmed [15:47:50] I think that should really be moved out of ExitModule [15:48:17] in theory it would be memory safe as they are all loaded in InitModule [15:51:18] but they are all definitely kept loaded and not released if you exit the simulation and go back into it without completely closing Orbiter [15:58:47] yeah, that's probably not good [16:07:11] morning! [16:11:45] good afternoon [16:12:17] what's up? [16:13:08] had to fix a bad RTCC bug. The usual :D [16:14:29] and "the bug which shall not be named" made an appearance too [16:16:04] SSV has cleaner code for fonts and brushes etc., using static class members. But I am not 100% sure this wouldn't cause issues if you have more than one Shuttle. [16:16:15] hmm [16:16:29] maybe if you had two Shuttles and you deleted one from the simulation [16:16:33] it haunts us forever :D [16:16:35] then the SSV code might do bad things :D [16:16:47] so not sure it's the right way in this case to steal from it [16:29:59] jarmonik definitely suggests creating fonts and such in clbkPostCreation [16:30:25] I doubt it has a big memory impact to have allocated them more than once if you have multiple vessels of the same type [16:33:07] if there is some other strange thing going on with the EMS scroll though then this wouldn't fix it [17:03:29] I feel like this is not a problem we've always had though. [17:04:10] maybe VC related? [17:05:01] yeah that would line up with what I remember [17:05:13] for reloading at least [17:05:25] the sm panel, I'm not sure [17:05:57] I don't think I remember it in 2020 when I started [17:12:57] hmm, Alex was talking about it on October 17 2020 [17:21:10] June 1, 2020: [17:21:21] [16:35:57] I had the weirdest bugs yesterday. In one case the LVDC IGM was freaking out, but it must have a separate cause, as the acceleration of the Saturn was too high? And suddenly I got a push and the stack fell quickly towards Earth [17:21:23] [16:36:09] and on the next attempt the SM was randomly missing a panel [17:21:24] [16:36:15] did a rebuild and that seemed to fix it [17:21:26] this is the first reference I see [17:22:20] interesting [17:24:29] indy91 I'm sure it's ccompletely unrelated, but how strange is it that you're testing a LVDC state vector issue, and get this bug just now [17:29:11] LM VC and ML mesh updates were merged in May [17:41:15] at least the LM VC can't be the cause directly as there is no LM yet when we get this bug during launch [17:41:31] it definitely is something uninitialized causing it [17:41:38] well, I am pretty sure at least [17:41:58] it typically happens on the first Orbiter start after a reboot [21:50:32] night! [16:40:16] good morning [16:41:01] morning! [17:28:48] hi guys [17:29:11] what's up? [17:29:59] hey [17:30:58] I think I figured out one more part of calculating constants for ropes for later launch years [17:31:06] oh nice! [17:31:13] not an ideal solution, but it will work fine [17:31:32] so I mostly have the LGC lunar ephemeris to figure out [17:32:01] and then we could give it a try to assemble a NBY 1973 rope [17:33:12] oh [17:33:17] LGC lunar ephemeris [17:33:18] LGC [17:33:31] I guess I already have everything for a NBY 1973 Artemis now [17:34:46] there is a bit more to do so that the padload generator can generate a padload for that [17:34:54] so not 100% there yet [17:36:43] instead of having to change code every time we create a new rope I want it to be able to choose a PIOS data set instead [17:36:57] so you would Artemis and it defaults to the NBY 72 data set [17:37:02] And then you can manually change that [17:37:17] so you would choose Artemis* [17:38:04] the function that creates the data for the ropes also outputs a data set which you can put in a text file with all data sets [17:38:33] so new ropes, that are functionally the same but valid for a different year, would then be supported without code changes [17:41:10] very nice [21:45:17] night! [15:54:59] hey [15:56:40] hi Nik [15:58:52] I reflew that -25s launch scenerio a few more times (4 I think), I never got a bad orbit again. [15:59:41] hmm [16:00:22] random issues like this are annoying [16:02:44] I wonder if It's related to the fact that I saved *at* -25 seconds. I usually pause before saving, but maybe I forgot [16:02:51] I could try to replicate [16:05:15] ah so it happened when you saved at T-25s and not when you, later, loaded that same scenario? [16:14:47] morning! [16:20:39] hey Mike [16:47:38] yeah, that's what I'm thinking. [16:52:39] any idea if you had gone from T-4h at that point or had you saved/loaded at all? [17:05:33] from -4. no saving [17:29:52] it would be nice to be able to replicate this [20:07:23] cya! [17:20:44] morning! [17:21:41] hey Mike [17:32:25] what's up? [17:37:34] just finished an update for the padload generator [17:37:48] it can now generate constants for ropes [17:37:57] nice! [17:38:15] not finished for the LGC, but it should be possible to modify Artemis now [17:38:27] are you going to keep going until you finish LGC, or switch to another project? [17:38:54] I can finish half of what remains for the LGC quickly, as we have an MIT memo on it [17:38:56] solar ephemeris [17:39:05] but the lunar ephemeris is a bit more tricky [17:39:08] need to figure that out [17:39:31] so I won't do it as my only project, but more like thinking about how I will do it [17:39:38] makes sense :D [17:45:15] because of 2000 being a leap year some simplified date calculations should not break down until much later [17:45:35] so I think everything, including the RTCC, would also work in 2024 [17:45:44] it doesn't allow 2024 as an input year right now [17:45:47] I need to change that [17:46:02] there are no ephemeris tape limitations we have, which was a reason to have a limit there [17:47:31] eventually my lunar inertial orientation subroutine numbers will break down because they are just linear projections from Artemis/Luminary 1E [17:47:42] but most of it seems periodic actually [17:47:56] the difference from AGC to Orbiter is smaller in 2000 than in 1972 [17:48:00] slightly [17:48:04] in between it's larger [17:48:08] so, periodic [17:48:11] interesting [17:48:33] I guess this is for the same reason why an unmodified Orbiter config file for the Moon is possible to be used [17:48:46] for the Moon rotations [17:51:18] you get errors on the magnitude of 200-300 meters on the surface. But this is corrected with a padload and was larger for the actual padloads than we have to use anyway. [18:39:57] looks like we might be getting proper soft-dock code in OO [19:21:25] oh that could be quite nice [19:21:35] bit of a headache, taking it easy. cya! [16:39:13] morning! [16:45:48] hey [16:56:37] hey guys [16:57:48] what's up? [17:00:05] annoyed by RTCC bugs I am finding :D [17:00:21] but first I will hopefully have a RTCC fix, then I will look at those [17:06:54] improperly shutting down weechat apparently [17:07:28] #debianproblems [17:27:48] I think I need to really dumb down how the RTCC environment change routine works [17:28:13] I was testing the map update PAD and wanted to compared with the sunrise/sunset display [17:28:17] that display is failing [17:29:17] shortly before LOI. It finds the sunset without trouble, but then it searches for the terminator sunset, so on the surface below the ground track [17:30:17] the environment change routine has more tasks to do than it currently can. Right now it uses some complex geometry to find sunset/sunrise more quickly [17:31:04] but I think I will replace that with... interval halfing :D [17:33:51] I don't know 100% if that's how that really worked, but I think the only way that routine can find a lot of different types of environment changes is by the dumbest, simplest search algorithm [17:36:21] hahaha [21:08:52] night! [15:26:34] good morning [15:30:08] hey Matt [15:37:46] I have not forgotten about your ASTP PR [15:38:19] thanks! I've got a RTCC fix PR coming up soon too... [15:38:50] I spent the last two evenings debugging a weird Orbiter issue for jarmonik, but that appears to be solved now [15:38:56] so back to NASSP [15:39:14] oh nice. Docking port related? [15:40:09] potential uninitialized elevation data [15:40:48] turned out to be something weirder, which I don't fully understand, but it seems to be fixed now [15:41:38] I did want to test the docking port feature, but that would require me to write vessel code to test [15:43:14] looking at the code, though I do think we can use the new MoveDock function to simulate docking index angle [15:44:01] with a bit of trickery I am sure it was always possible. Create the docking port at the specific angle if the CSM is at the right attitude. [15:44:06] actually [15:44:19] which one should be changed. CSM or LM docking port :D [15:44:47] I guess it has to be CSM. The docking probe class is in the CSM and is really the active code in this. [15:45:45] in my ASTP branch I managed to get CSM + Skylab CG calculations working for the DAP PAD with the input docking angle [15:46:06] but what really should be changed there is the (currently hardcoded) nominal docking angle of 240° [15:46:11] or minus 60° really [15:46:33] was hardcoded like this in the real RTCC, but got very likely changed to a system parameter for Skylab [15:46:40] easily changed [15:52:11] oh and the RTE is super confusing with this. It says it has 60° as the docking angle input. But all other processors use a delta docking angle, so difference to the nominal one... [15:52:20] maybe it's an error in the RTE MED description [15:54:56] nope, seems like you had to input 62° in the RTE for a delta docking angle of 2° [15:55:00] or 58°, who knows... [15:55:39] thats kinda odd [15:56:18] yeah [15:56:45] it defaults to a system parameter of 60°. It's nice knowing the name of that parameter. [15:57:17] I think I will try to be consistent. That system parameter will be loadable in the RTCC mission constants file. [15:57:28] And all input, including the RTE, will use a delta docking angle [15:57:42] MPT stores a delta docking angle, that is very clear from the documentation [15:59:04] when the Apollo 11 RETRO was a bit confused about the lunar entry REFSMMAT, he talked a bit to the FIDO about LVLH angle transformation conventions etc. [15:59:25] It definitely seems like different departments (e.g. FIDO, RETRO, GUIDO) all wrote separate requirements for the RTCC [15:59:45] and that's how you might end up with some inconsistencies like this [16:02:55] yeah, I can imagine [16:05:40] it doesn't really make sense for the RTE to do this. [16:05:50] RTE uses the CSM/LM burn simulation. [16:06:00] that in turn uses the CG calculation function [16:06:10] and that function uses delta docking angle as input... [16:28:43] I still have this instinct to be careful what I am typing when Orbiter is open in the background [16:28:50] will take a while to get rid of that :D [16:29:24] morning! [16:29:27] lol [16:29:32] hey haha [16:32:14] oh yeah definitely haha [17:53:58] hmm [17:54:10] thewonderidiot, a memo written on April 20, 1971. [17:54:17] that doesn't make it into Luminary 1E, does it :D [17:54:59] might explain why I didn't match the controlled constants exactly haha [17:55:19] it's an updated memo. I bet if I use the original, inferior one the values will match. [17:56:42] hmm but in that case it should have two identical values for something, but it doesn't. So maybe the logic behind the updated memo should be in the constants and I just got something wrong... [17:57:26] "Release of Luminary 1E for Rope Manufacture, Rev 210 ........ 3/22/71" [17:57:32] yeah, definitely not :D [18:19:18] "The constants LOS O, LOSR, C, OMEGAC and PHASE [18:19:21] C [18:19:21] have traditionally been [18:19:22] determined empirically by curve-fitting data from the JPL Ephemeris Tapes over [18:19:22] specified time intervals." [18:19:35] I bet that was still the case for Luminary 1E [18:19:52] the memo just presented a way with some orbital elements calculations to get something similar [18:20:02] oh interesting [18:20:33] which then might or might not have been used for Skylark [18:52:30] LGC solar ephemeris is go [18:52:38] nice! [18:52:51] accuracy 0.03° at worst, I don't think you will get anything better with this simplified scheme [18:53:14] on 13 they even gave a more accurate unit vector to the crew I believe [18:53:22] instead of using the onboard sun calculation [18:55:40] for the luna ephemeris I might end up doing curve fitting [18:55:44] lunar* [18:55:54] more complicated, but not the end of the world [21:25:09] night! [13:26:03] hello [13:27:48] hey [13:40:07] I've added something to the ASTP scenario for the requred addons [13:40:41] what have you tested so far? Let the S-IVB do a bunch of maneuvering? [13:54:46] I flew a bit of the new skylab and astp scenerios, and I had sent some commands through thr new PAMFD interface [13:56:03] ah nice. Yeah anything regarding the LVDC orbital guidance has the best chance of more bugs. Most other things in the PR are pretty safe or have been tested a bunch. [13:58:58] In the PAMFD I ran out of new buttons for the new commands, so, that's why I changed it to the way it is now [14:10:05] back in a few hours [17:25:34] back [17:25:59] morning! [17:29:10] hey [17:29:43] what's up? [17:30:40] was going to look at some more RTCC bugs. And you? [17:30:57] I don't know haha [17:31:00] trying to decide what to do [17:32:51] ah yeah, that time after a multi month project :D [17:33:02] yep haha [17:36:20] Skylark almost seems like the crowning achievement with the AGC. What can really top that. [17:37:01] yeah I know [17:37:19] all of that work kind of culminated in Skylark [17:38:04] seeing V79E do something will be very nice, definitely something to look forward to. [17:38:45] I don't think I have that much desired to fly Apollo missions to the Moon past 1974, but it's definitely a good capability to have, calculating the rope constrants. [17:38:51] desire* [17:39:02] so I definitely want to finish that [17:39:24] yeah for sure [17:39:49] and yeah. there's still Comanche 67 to be disassembled, and Comanche 72 and Manche72 rev 3 to be reconstructed [17:40:27] I should get back to working on Sunspot too. Skylark totally derailed that work [17:40:40] the RTCC is an endless stream of new projects, but also has diminishing returns. In a bunch of cases I would like to rewrite some processors from scratch, with very little additional capabilities as the end result. Not as fun as starting something completely new. [17:41:12] Yeah Sunspot. Too bad we never got that GSOP from UHCL... [17:41:19] yeah :( [17:41:23] was probably too late for that anyway when we were thinking about it [17:41:36] I hadn't requested anything in a while [17:41:44] not sure when they stopped taking requests [17:42:38] Sunspot would probably have a lot of bugs, but, GSOP section 4 shows just how different it is. Would be quite fun for Earth orbital missions. [17:43:19] yeah. but we'll need a second source of sunspot modules before that becomes possible [17:43:24] yep [17:43:40] so not as much stress getting the Block I branch up to date :D [17:43:48] hahaha indeed [17:43:55] I'd have to make the programer optional, too [17:45:29] when in doubt for me there's also always punch cards to go back to [17:46:21] Yeah there are so many non-RTCC NASSP projects I could do, too [17:46:50] moving system initialization to a later time in scenario loading will make Ryan very happen, because then we can load mission specific systems [17:47:06] speaking of, back when I convinced Ron to use a punchcard-like format for the AS-512 and AS-513 listings, I promised I would go back and convert the AS-206RAM to the same format.... and still haven't yet haha [17:47:10] makes the separate J-type mission branch obsolete [17:47:17] oh yes, that would be a huge win for NASSP [17:48:06] converting AS-206RAM into punch cards sounds like a bunch of work, but not too difficult [17:48:08] also converting the checklists from XLS to TSV [17:48:57] and for me, converting our panels into dynamic classes where switches can be loaded individually based on mission [17:49:21] I really need to force Jordan to make that VC branch ready to be merged. Making big changes like that is really not going to work with that branch not merged. [17:50:12] that VC branch still doesn't show a diff on Github... [17:50:26] how am I ever going to approve that if I can't see a diff [17:50:52] it might be the mesh changes, which are processed as code, more than anything else. [17:51:07] meshes being text files essentially [17:51:14] probably huge amount of changes there [17:52:12] yeah, massive PRs like that are difficult to manage [17:53:12] maybe I can tell git to ignore mesh files as code, just for showing a diff [17:53:21] the actual code changes should be managable [17:53:29] it's also a well tested branch, so, no big deal there [17:53:40] I just want to see the code changes somehow haha [17:55:20] when in doubt you can always check out his branch locally and use `git diff` on specific folders [17:55:36] ah yeah, that will probably be the way [17:55:40] Orbitersdk folder [17:57:18] I think git diff has some options there, so, I'm sure it is possible to see [17:58:27] yeah, check out his branch and then do [17:58:34] git diff Orbiter2016 Orbitersdk [17:58:40] should do the trick [17:59:37] seeing that branch name does remind me -- they're threatening to make an OpenOrbiter release this year aren't they? [17:59:55] when do we start deciding what officially goes into NASSP 8? haha [18:01:41] first we have to decide if NASSP 8 is for Open Orbiter or the Beta :D [18:01:47] and I tending towards OO now [18:02:47] yeah, OO seems like it's starting to be in pretty good shape [18:02:51] if we decide for OO, not sure how to handle the switch. if we should try to release a stable last release for the Beta of some kind. [18:03:22] otherwise there was never a NASSP release for Orbiter 2016 or the Beta... [18:03:54] somehow the fact that there was never even an Orbiter 2016 release escaped my brain haha [18:04:00] right, NASSP 7 was 2010... [18:04:56] to be fairl, there are reasons for that [18:05:24] Apollo 9 does not work in Orbiter 2016, because of the docked moments of inertia being calculated wrong [18:05:44] and as the idea for NASSP 8 was to include Apollo 7-11, that basically means Orbiter 2016 is out [18:05:52] as a release version [18:06:08] right [18:06:14] Orbiter Beta is definitely more feasible, but even there are annoying issues [18:06:29] like not properly supporting two vessels with two different VCs in the same simulation [18:07:07] that's why, when you load a scenario with both CSM and LM, you have to change the view point once before clickspots for switches work [18:07:18] I am sure we could figure out a workaround, but, it's fixed in OO [18:08:31] the installation process with the Orbiter Beta is of course also quite annoying [18:09:12] yes, it really is haha [18:09:13] I was always expecting one last Orbiter release from martins which we would develop NASSP for, but, I guess that now is Open Orbiter 2024 or whatever it will be called [18:09:53] so it probably makes the most sense to aim for that [18:10:26] and then we do a hard cut and say "no additional Orbiter features or bugfixes are worth waiting" and should really try to release NASSP 8 for that [18:10:41] will still take years probably, but, might be the best path forward [18:11:14] yep [18:55:04] hello [18:56:58] indy91, my vode for 8.0 release is with OO [18:57:07] whatever "release" means [18:58:20] yeah, whatever release means... we have sort of transitioned to a infinite development cycle. I think it's still worth changing up a bit how we develop if a stable OO version is available. [18:59:16] removing code that we still have for backwards compatibility, trying to really make a push to delete redundant code etc. [19:00:58] I guess I should try our OO branch some time... might help with getting the main NASSP development branch to OO :D [19:01:51] Within our userbase there are probably people that would be fine updating twice per year, and people who want the latest version. Judging by the forum and discord, it would seem to be that there are more more people in group #2, but I feel like most of the people that aren't on discord and the forum are in group #1 [19:05:07] twice per year? my NASSP install from early 2019 is still working fine, thank you very much :D [19:06:32] Group 1, Group 2, Mike [19:07:11] oh and let's not forget Turry's friend Alejandro who was actually waiting for a NASSP release that includes lunar landings again until just a few weeks ago. [19:07:15] was still using NASSP 6 [19:07:35] I was just about to comment about that haha [19:07:38] hahahah [19:07:40] 6.4.3 [19:07:48] "Last stable version" [19:09:08] I never had much luck with that one. the default Orbiter tools are fine for situations were you have like 50km/sec of DV, not so much for Apollo. [19:12:24] I almost got to Mars with the default tools, but then Mars jumped out of the way. Open Orbiter 64bit. But back in the day, I don't even remember how much interplanetary flight I tried... [19:12:46] I know I used NASSP in 2007, I had found some printed out checklists. But that was probably Apollo 7. [19:13:17] I've been mostly messing around with orbital rendezvous before I became part of NASSP [19:14:15] probably not in 2007 though, didn't know much about orbital mechanics yet [19:14:51] stuff like STS Payloads is what I used a bunch [19:22:26] I used to do a lot of TransX flights out to Jupiter and Saturn in the O2006 era. For Mars you kinda need to know what you're doing. [19:23:31] I did a lot of station-building in earth orbit [20:19:06] cya! [13:36:16] hey [15:16:10] good morning [15:31:00] morning! [15:32:26] hello [15:38:32] hey guys [15:38:39] indy91 your PR looks simple enough, but I don't quite understand what's going on there [15:43:59] ultimately the bug was an inconsistent set of masses given to the CSM/LM burn simulation [15:44:19] total mass was correct, CSM and LM masses separately was not in all cases [15:44:35] that lead to wrong SPS trims being calculated [15:46:22] hmm [15:46:34] actually, I think it's the inertial burn attitude that was wrong [15:46:45] due to LM mass not taken into account separately [15:46:57] SPS gimbal trim angle was calculated correctly [15:47:01] angles* [15:47:13] and those trims were used for the TEI REFSMMAT calculation [15:47:23] which then didn't match the burn attitude [15:47:36] so you didn't get exactly 180,0,0 with the TEI REFSMMAT [15:48:53] so essentially the function that calculated CG location was not supplied with the correct CSM and LM masses separately [15:50:22] there also is the issue that only the early version of the CSM/LM burn sim really tracked LM ascent and descent stage masses separate I believe. [15:50:40] to simulate something like the Apollo 5 fire in the hole [15:51:04] later RTCC functions that use the burn sim don't have that separate [15:51:29] -integin.LMAWT = 0.0; [15:51:31] +integin.LMDWT = vars->LMWeight; [15:51:32] +integin.LMAWT = vars->LMWeight; [15:51:33] +integin.LMDWT = 0.0; [15:51:34] uhh [15:51:39] first two minus :D [15:52:06] if this looks weird, inputting the toal LM weight as the ascent stage weight just makes sure it is always being used for the LM weight [15:52:16] no matter the configuration (ascent or full LM) [15:53:05] ahh okay [15:54:21] they likely removed the separate ascent/descent weight inputs in later versions, because no function that uses it in the later RTCC even tracks that separately. It just controls it with a configuration code (ascent or full LM) [15:56:11] a common issue I have where I have proper documentation for two RTCC functions, one of which calls the other. But they aren't the same revisions of the functions so you get interface differences. [15:56:29] A bit like stiching together Sundance with rope modules of different revisions :D [16:03:58] :P [16:07:09] the comparison works pretty well. I also have three main RTCC revisions, Earth Orbit Rendezvous Program (EORP), Lunar Mission Simulation Program (LMSP), Lunar Landing Program (LLP) [16:07:32] all I want is LLP, but the reality is I have 30% EORP, 60% LMSP and 10% LLP [16:08:32] and that 100% is only like 25% of the whole RTCC... [16:15:19] wow yeah that comparison works very well :D [16:22:12] I mean there is slightly more to it. There were steps before EORP. I have some flowcharts for unmanned missions as well. Like simulating the calculations that Solarium does for burn attitude, for example. [16:22:36] can't do much with that, but, we have it [19:09:06] // Continue writing until [19:09:10] thanks, past me :D [19:09:15] until what?? [19:35:48] forever [16:26:03] morning! [16:27:11] hey [16:56:35] what's up? [16:59:37] hey guys [16:59:59] I think it's time for a long anticipated update. Apollo 12 MCC lunar orbit map updates lol [17:11:14] finially [17:15:04] oh man I have been waiting years for that :D [17:15:27] finally the mission becomes usable [17:15:35] people have been lost without AOS/LOS times for years [21:34:22] that was a lot of MCC states to add at once... [21:34:34] now my head is spinning haha. Night! [16:00:05] hey [16:14:29] hey Nik [16:35:28] morning! [16:47:07] OpenOrbiter is making me feel better about our memory issues [16:50:37] hey guys [16:50:54] seeing Apollo 11 being flown all in one session by Turry is what made me feel really good in that regard :D [16:54:56] I like the newly established tradition of finding and fixing bugs related to never reloading before he even lands [16:56:17] kind of a hard test cast to replicated on my end haha [17:00:00] although for telemetry I will probably let it run for a long time to make sure my buffers are safe [17:02:08] the bug he had on Apollo 10 is something I should have spotted in code much earlier, but just from a testing perspective, a bit unlikely to be found :D [17:02:37] a pointer to the LM, which is used for uplinking, was only set on a scenario reload when the LM is present in the simuation [17:03:09] so the only way to get the CTD from it was to fly from before CSM sep after TLI to the first LGC uplink in lunar orbit without reload [17:08:53] I don't have that kind of time haha [17:11:19] I am curious how long it takes the average user to complete a mission though. [17:25:30] probably depends on the experience level [17:28:15] in my experience, at least 9 years [17:31:52] ever heard of lunar ascent? :D [17:32:09] nope! [17:32:30] but you pressed the abort or abort stage button before, right? [17:32:36] that's true, I have [17:32:44] so I guess I have heard of it [17:32:49] and then said it would take too long to demonstrate all that [17:33:03] pretty much haha [17:39:23] not to spoil the ending, but there is a second landing you get to do if you fly far enough into the mission [17:40:28] oh man [17:43:40] it's right before the water level [17:44:40] oh no, not a water level! [17:45:45] the fire level right before is pretty nice though [17:47:34] true [19:18:19] I have a terrible secret [19:18:27] I usually really like water levels [19:43:12] Well that settels it then. I guess you'll have to fly a full mission :) [19:43:26] I think the fear of water levels comes from clunky 90s swimming controls :D [19:44:12] some of those were rather hard [21:10:35] night! [16:01:44] good afternoon [16:39:44] morning! [17:12:57] hi Mike [17:13:12] what's up? [17:20:04] not too much [17:21:41] I might circle back on our PowerMerge code. I gave my electrical engineer friend a hand-sketch of our "power from two busses" situation, and he handed me back current draw for each source in about 30 seconds [17:25:10] oh nice! [17:43:14] that and doing some long-term testing for OO [18:00:21] I've been falling down the fractal projects rabbit hole [18:01:41] I need to make a board to interface with a CDU -> I want to be able to hook up an FPGA AGC to this board -> I want to be able to put the AGC and the Monitor into the same FPGA on a cheap dev board -> the Monitor is doing things that don't quite align with the user manual, I should fix them [18:12:57] ahh fun [15:58:43] hey [16:12:57] hey Nik [16:15:19] what's up? [16:19:46] oh, still trying to help OO along to a release [16:20:00] and flying Apollo 11 in between [16:21:54] ah nice [16:22:11] memory leak found yet? [16:23:36] I think so. [16:25:15] I think it was mostly memory fragmentation, rather than a true leak, at least with the last few commits before the fix. [16:35:32] interesting [16:51:30] drift rates have been reasonable [16:52:24] I think I will probably just ommit them from Apollo 12 for now, until we have the correct CMC software [16:56:50] morning! [17:03:27] hi Mike [17:18:15] hey Mike [17:39:55] n7275, are you reading what I am writing on Discord about the LM ECS water separator? I think it's reaching too many RPM due to a change in your big systems update. [17:45:21] oh, I hadn't looked at Discord yet [20:16:08] cya! [15:24:21] good morning [15:30:45] hey Matt [15:34:25] morning! [15:36:33] good afternoon [16:14:45] what's up? [16:16:40] Apollo 12 LM activation, for testing of the added MCC map updates [16:16:53] the timing of those updates is pretty nice [16:17:16] the RTCC doesn't need to do any additional state vector propagation, just "search from current time for AOS/LOS etc." [16:17:43] so on the RTCC side all of the map updates are exactly identical. Just 15 lines of code for all of them combined. [16:18:09] on the MCC side of course, scheduling all the updates, it was a lot of additions :D [16:19:17] very nice :D [16:19:27] and that's what I need to test that it all still works [21:05:35] night! [17:39:09] morning! [17:49:54] hey Mike [15:05:56] hello [15:23:20] hello [15:24:22] winter has made a bit of a return, taking out our internet. my regular IRC connection will be a bit spotty for a few days [15:36:49] hey Matt [15:37:02] hi Nik [15:37:22] if it was like 5° colder here then we would have the same, because it's raining all day [15:38:40] there's about a 5cm thick layer of ice on all the trees, powerlines, and apparently phone/internet [15:38:56] ice, ouch [15:39:08] yeah that's going to be annoying for infrastructure [15:39:33] yeah, like half the state has no power [15:41:03] then spotty Internet isn't the worst thing in the world, comparativelx [15:41:06] y [15:44:04] yeah we have power, so I'm glad for that. the power for my building comes from a hydro plant that's only like 150ft away [15:44:36] that sounds nice, but is that actually feeding the power directly like that :D [15:47:41] it's a bit more complicated, but the building I live in is an old textiles mill that used to get its power from that plant. so there's a substation in between so it can connect to the grid [15:47:50] aaah nice [15:48:02] got your own transformer [15:48:40] winter is usually not a problem for power here, most power lines are underground. I imagine with 5 cm of ice though some of the larger overhead power lines might start get into trouble though [15:49:14] on the hottest days of summer we had a few shorter power outage in the last years [15:49:18] that is more annoying [15:49:51] there was a bit ice storm in the late 90s that everyone still remembers a "The Ice storm of '98" [15:50:43] it completely destroyed a bunch of transmission lines [15:51:05] especially the ones coming down from Quebec [15:52:16] "big" not "bit" [15:53:37] actual ice storms are very rare here, luckily [15:54:08] our big winter storms are usually a problem because of the storm, not because of snow or ice [15:57:25] I recall some really bad flooding a few years ago [15:57:39] hearing about it in the news [16:01:00] oh, apparently I'm back [16:15:38] one version of you is back, yes, haha [16:17:18] it's the real me, here I'll prove it. [16:17:53] the IMU drift, making the IMU less accurate than GDC/EMS/ASA is making me want to add drift and biases to those too [16:22:02] true [16:22:06] especially the GDC [16:22:29] our AGS calibration method already seems to be inaccurate enough to give (small) biases :D [16:23:22] our GDC at least degrades near gimbal lock already [16:26:16] morning! [16:26:19] hey Mike [16:26:20] I like the sound of making all of these things worse :D [16:26:29] I can't even get the IU state vector under control without it... [16:26:58] damn venting or whatever causes it [16:27:03] who had the idea to add that! [16:27:16] vent bent ascent, lament [16:27:29] lol [16:31:04] yeah it's definitely the ascent, whatever problem it is, it happens somewhere from liftoff to shortly after insertion [16:32:04] in that case it could only really be venting if it happens right after the accelerometers are not used anymore [16:36:04] that seems like a lot to be just venting though [16:36:45] it would be a transient of some sort [16:37:12] but I don't really believe in it. It's something else with the acceleration measurement. [16:37:20] and I don't ever get it myself [16:39:07] I was only able to get it once [16:39:26] I have not been able to repeat that [21:42:56] night! [16:10:20] morning! [16:12:05] hey [16:12:19] what's up? [16:12:42] continuing with Apollo 12. And you? [16:13:13] continuing to be confused by CDUs :D [16:13:56] ah of course [16:16:35] researching the behavior again that leads to 1202s? [16:26:54] indirectly -- I'm mostly working towards powering one on [16:28:43] but last night James and I stumbled on a confusing input to the D/A module that is mentioned in almost no documents [16:29:09] oh [16:29:26] anything I could have heard about? [16:30:18] it's an input that inhibits coarse error from going into the mixing amplifier except during CDU zero... or something like that [16:31:23] usually drawing 2015566, IMU-Rend Radar CDU Block Diagram - Block II (2 Line Mech) is really useful in showing internal connections like that. but the latest revision NARA has, rev B, does not show it [16:31:38] they also have the original revision -, which is too early to be of use [16:32:45] so last night I noticed that the revision history block of rev B says something incredibly strange [16:32:53] https://i.imgur.com/7cv4qFY.png [16:33:16] "Redrawn per TDRR 17147, which states that Rev A was replaced. Actually Rev A was destroyed when it was in the Rev - configuration." [16:33:57] which probably explains why NARA doesn't have rev A [16:34:49] could this be a RR CDU only thing? [16:34:54] but miraculously, Rev A actually has come down to us anyways from Don [16:35:04] https://www.ibiblio.org/apollo/Documents/blk2_cm_lem_mechanization_diagrams.pdf pdf page 14 [16:35:17] or I guess everything except RR CDU [16:35:34] its revision block just says "Redrawn per TDRR" without giving a number, so presumably one wasn't ever assigned [16:35:53] and it shows the input, and does not specify that it is only meant for RR or only for IMU [16:36:19] so, this feature was a late-ish addition [16:36:33] for it to be in Rev A, which was destroyed and never released, and then not in Rev B is very peculiar [16:37:09] hmm [16:37:13] yeah confusing [16:39:57] classic CDU revision history madness [16:41:50] "should have just stuck with the Block I mechanical ones" - CaptainSwag101 [16:42:59] having implemented that in the Block I branch, aside from the special reentry mode it translates to code fairly well :D [16:45:08] hehehe [17:02:57] so long story short, I'm pretty confident that the D/A and EC&L modules have this hardware [17:03:08] whether it's actually connected across the backplane, I will have to measure [17:03:27] because the documentation definitely will not help answer that :P [17:17:20] results maybe surprise. You have a bit of a weird CDU :D [17:17:25] may* [17:17:53] ok, the category of weird CDUs is probably all CDUs. But you know what I mean. [17:18:41] hehehehe [21:49:04] night! [15:57:57] good morning [16:08:01] hey Nik [16:10:13] hey [16:10:32] how is the weather situation? [16:44:22] morning! [16:51:00] getting better, the ice has mostly melted but all the trees look like someone turned the gravity up on them for a few days [16:53:00] some of these pine trees aparently aren't the "all season" model [16:55:08] https://www.pressherald.com/2024/03/25/restoration-continues-as-tens-of-thousands-of-mainers-remain-without-power/ [16:59:12] hey Mike [16:59:35] what's up? [16:59:41] oh hi Mike [17:00:02] poor trees, but I guess most of them will recover [17:00:21] they make new ones every day :) [17:01:37] continuing with Apollo 12 [17:01:56] I've now made use of a weird feature of Orbiter annotations, which previously just got in my way [17:02:12] if you have two empty spaces in a row the annotation starts a new line [17:02:28] I have a fairly complicated PAD, which is only used once [17:02:41] so I don't want to add all the code required for a special PAD type [17:02:58] so instead I am using GENERICPAD, which is just a char array [17:03:29] The only good way to have multiple displayed lines with that is this double empty space feature [17:03:54] otherwise the PAD string would be saved to scenarios with \n (new line) as well [17:03:55] ahh that does sound useful [17:03:57] which doesn't load right [17:04:32] slight downside, in the MCCPreAdvisoryData.log file it shows up with the double empty space [17:04:41] so all one line [17:04:49] but that's not so bad I think [17:04:56] it's only this one PAD anyway [17:05:33] our MCC was written by Dan more in C than C++ [17:05:53] yeah as long as scanf doesn't decide to ignore them anywhere that should be fine [17:05:59] what I would kind of prefer for the generic PAD would be a std::vector of std::string [17:06:40] but that is contrary to how PADs work in MCC [17:06:51] there is a "calloc" [17:07:04] padForm = calloc(1, sizeof(GENERICPAD)); [17:07:17] I don't think C++ vectors or strings work with this :D [17:07:47] ahh no [17:08:20] so thanks to the double empty space feature I don't have to totally rewrite this for multi line, generic PADs [17:08:44] I used to get these by accident with block data [17:08:55] when not all 8 sets of block data were used [17:09:09] then I had two empty spaces as well [17:09:40] I hope it wasn't this that used to cause CTDs... [17:09:57] Orbiter pretty much exclusively uses char* strings internally any way. I think converting it(and the api) to std::string is on someone's longterm list [17:10:28] block data used to occasionally cause CTDs. I could never really figure out how, it wasn't the actual calculation, it was the PAD. I just made it memory safer and memory safer and one day it was gone. [17:13:20] yeah I studied the Orbiter code for this a bit, to understand if I could really use this double empty space [17:13:31] I haven't seen it documented, but it seems intentional lol [17:14:18] if I didn't use it, annotations automatically start a new line when the annotation box is exceeded on the right, that's what I had at first [17:15:54] I think this will be a good system for PADs that only used once or twice [19:58:14] cya! [15:39:58] hey [16:38:30] morning! [16:45:10] hey Mike [16:49:26] any luck with the mystery inhibit? [17:01:31] it is totally unwired for the RR [17:01:47] which makes sense, because the coarse error is also not wired to the D/A module, so there is nothing to inhibit [17:02:19] and at least in a LM CDU, it is wired as the module schematics indicate for the IMU [17:02:35] er, I mean for LM IMU [17:02:45] I haven't checked the CM IMU case but I can't imagine that it's different [17:03:19] most of the IMU differences between LM and CM CDUs are tied to the additional functionalities like AGS interfacing or Saturn takeover [17:05:32] yeah, that all sounds logical [17:08:23] my new theory is that it is related to the IMU turn-on mode in the CDU, which ended up not being used [17:09:03] it makes a lot of sense in that context, and also explains why this inhibit is not really mentioned by documentation [17:12:11] what would that mode have done? [17:12:54] it does a coarse align while the read counter is held at 0 [17:17:23] well sort of? [17:17:33] it also sets the 90 degree bit in the read counter [17:20:29] here: https://i.imgur.com/HCIN2Tz.png [17:24:12] ah, classic CDU [17:36:28] yeah lol [17:36:59] James and I had already been confused by and beeped out the turn-on enable signal Q [17:38:09] he has a theory is that that wire from the unused turn-on feature was the reason the 90 degree bit specifically was flipped in a read counter in the Apollo 12 lightning strike [17:38:47] because that wire has the ability to set the 90 degree bit in the read counter, and has a long backplane wire associated with it, that might act as an antenna -- which is true of no other read counter bits [17:42:24] oh wow haha [17:48:06] sounds like a good theory imo [17:48:32] the postflight documentation said they tested the tumbling condition, but maybe by artificially setting the bit [18:00:02] hey [18:01:11] yo [18:19:32] what's up? [18:33:55] lots of CDU research and planning [18:34:05] which has taken diversions into PSA for a bit [18:40:44] the PSA is interesting. from what I can tell from the drawings, they didn't use the gardner-denver automatic wire wrap machines for them [18:40:54] all of the wire-wrapping in the backplane was done by hand [18:41:23] and the side effect of that... is that the card decks for the machines required all nets to be given <=8 character names [18:41:47] which is not a limitation for the PSA.... and so they didn't actually bother to even give PSA backplane wires names at all [18:45:35] hey Matt [18:45:51] wiring too simple to be properly documented, making wiring (60 years later) complicated [18:52:34] the upside is that we actually have the complete list of all backplane wires from NARA https://archive.org/details/apertureCardBox471NARASW_images/page/n1347/mode/1up?view=theater [18:52:49] there's just no names, so it's a bit difficult to parse and find the connection you're looking for lol [19:00:31] https://www.abcund123.de/wp-content/uploads/2013/11/Spuren.png [19:32:13] hahahaha yes [21:34:01] night! [13:38:32] hello [13:46:12] hey Matt [13:50:34] have you seen the documents that DAN posted on Discord? [13:52:12] I have not yet. let me check [13:52:36] yeah are all about telemetry processing [13:52:40] well, mostly [13:55:40] oh my. this will be very useful [14:04:21] yeah it's a great find [14:04:33] not what I was expecting, I thought there was more about displays [15:54:25] morning! [15:57:15] hey Mike [15:58:03] what's up? [16:06:39] hey [16:07:10] thinking about what I have to add to the MCC for Apollo 12s last day in lunar orbit [17:24:06] oh man, I'm definitely implimenting this calibration tape format [21:46:14] night! [14:20:34] hey [16:34:28] morning! [16:39:49] hey Mike [16:58:29] what's up? [16:59:58] I found another MOCR display format! [17:00:05] the mission plan table itself [17:00:50] https://www.flickr.com/photos/fori/4431078349/in/pool-1391010@N22/ [17:03:32] oh man, nice! [17:03:42] a good couple of days for MOCR displays haha [17:04:14] yep [17:04:36] I might finish the Apollo 12 MCC update today, so I can get to it [17:05:16] 6 hours to TEI with a bunch of activity, but I am basically through when I have tested the pre TEI map update [17:06:40] awesome :D [17:08:54] are you still working on your understand of CDUs? [17:09:04] understanding* [17:10:01] yep [17:10:32] I have one cracked open right now, and am working on resolving all of my uncertain pins for LM CDUs [17:11:56] also starting to go through my electrical engineering textbooks from university again [17:12:18] I have forgotten a lot and need to re-learn some things for the section of PSA I'm planning on rebuilding haha [17:16:49] yeah that doesn't sound like a simple job [17:17:20] I haven't worked with AC circuits in a loooong time haha [17:20:53] I have documentation of several versions of the RTCC function that calculates many parameters for the MPT display. This display clearly shows something that was in the (earlier) RTCC Lunar Mission Simulation Program, but not anymore in the (later) Lunar Landing Program. [17:21:16] I wonder if that was an accidental omission [17:21:59] or well, potentially this code could have been moved into the sub-function that gets called [17:22:32] apogee and perigee heights calculation [17:23:03] this "sub-function" works only with the AEG, or at least the version I have documentation for. That gives it an eccentricity limit of 0.85 max. [17:23:23] in the LMSP version of the MPT calculation it has some extra logic to handle larger eccentricities [17:23:35] like, still calculating a periapsis height even if e > 1.0 [17:24:03] LLP version does not have that, so I had assumed it wouldn't calculate anything [17:24:16] but this MPT display clearly has a HP (no HA) for pre LOI [17:24:53] is this definitely original, or is it possible that this was pieced together by the MOCR restoration team and these are mistakes they made? [17:24:59] it does appear to be a picture from that [17:25:18] I am 95% sure it is original [17:25:22] it just fits too exact [17:25:38] it has a maneuver on it that the FIDO used just to calculate a REFSMMAT [17:26:27] the places where HA and HP have a valid value vs. not (ZZZZZZ) fit exactly with the LMSP calculation [17:27:25] my best bet is that they changed this apsides heights determination subroutine to also handle cases where the AEG wouldn't work, highly elliptic or beyond [17:27:52] simply moved the code into it that was previously in the calling function [17:29:06] if I do it that way then I don't have to contradict any IBM RTCC documentation I have [17:29:25] nice :D [17:29:29] I just don't have the LLP version of the apsides heights determination subroutine [17:29:36] where this would have been moved [17:29:46] but I think it makes sense [17:29:56] expand the capabilities of that subroutine to all orbits [20:49:38] night! [13:37:22] hello [16:28:10] morning! [16:35:33] hey [16:35:55] hey hey [16:36:04] what's up? [16:36:47] tweaking the MPT display [16:37:22] of course using an Apollo 11 pre MCC-2 scenario [16:37:30] for best comparison with that real display [16:37:55] I wonder where the HA and HP for MCC-2 are coming from, those look like nonsense values haha [16:38:38] 167.3 x 3892.0 ??? [16:38:55] not exactly the actual orbit after MCC-2 [16:38:59] hahaha not quite [16:39:16] conically I am getting something like 41.0 x 99999.9 (display limit) [16:39:37] definitelly too eccentric for an AEG solution [16:40:05] but conic doesn't make sense either [16:40:28] and I don't think they would use the integrator to find an actual apoapsis/periapsis [16:46:10] which these heights wouldn't be anyway [16:49:12] but I am not worried about this, the values don't mean anything [16:56:14] https://www.flickr.com/photos/fori/4431078349/in/pool-1391010@N22/ [16:56:57] comparison imminent, copy and paste failure... [16:57:51] https://i.imgur.com/1edMxpd.png [16:57:56] still some formatting to do [16:58:00] but getting closer [16:59:02] example how it looked before: [16:59:03] https://cdn.discordapp.com/attachments/809545464331370506/1223680005427429418/image.png?ex=661abbd0&is=660846d0&hm=3a4f06657d55fd63f63a9c87beee157e74e8bd059861a0e2f3fb95d67ed717f1& [17:00:21] nice, looking good! [17:01:52] I am always running into trouble with the format being 1:1 [17:02:02] I can get rid of one letter from the maneuver code [17:02:07] I have one too many [17:03:17] hey guys [17:04:25] hey Matt [17:09:39] how did you find that mpt image? [17:13:47] I googled "mission plan table" :D [17:14:16] I think from MOCR 2, but before it was restored [17:14:21] photo is from 2010 [17:14:52] other than the displays from the restored MOCR this one seems to be original [17:14:59] different than* [17:16:44] cool [17:17:40] from the actual Apollo 11 mission [17:17:42] so not a sim or so [17:18:03] I can find everything about all the maneuvers on here from the FIDO transcript [17:18:23] probably even the exact time when this hardcopy was saved [17:18:51] the last maneuver on it was to calculate a REFSMMAT for the (not performed) lunar orbit plane change [17:18:59] that wasn't on the MPT for that long, I think [17:19:35] a few hours later they would have an ascent on the MPT most of the time, but not here yet [20:25:55] hmm [20:26:14] while JSC History has become less cooperative, NASA HQ might have some brewing [20:26:41] https://history2.nasa.gov/finding_aids.html [20:27:44] there are large Excel files, with some already scanned documents, listed [20:28:13] I can't access it directly, but one thing I see is e.g. an Apollo 10 SCOT [20:28:21] which we have, but, there will be more like it [20:34:45] a lot of the changes seem to be fresh, so, more things might be happening [20:34:58] and I definitely have a few PDF file names I could ask them about [20:35:14] Apollo 7 technical debriefing would be interesting :D [20:40:12] oooh interesting! [20:40:53] https://history2.nasa.gov/rg_finding_aids/rg_11_spaceflight_march2020.xlsx [20:40:59] best one I have seen so far [20:41:09] that has the PDF numbers of those documents [20:41:30] > ASTP CS/CM & DM System Handbook (Aug. 1, 1974) [20:43:18] wait, why am I not find that [20:43:23] Excel you are failing me [20:43:36] it's in a PDF, not in excel :P [20:43:56] oh [20:44:06] I was looking in that file I just linked :D [20:44:12] where did you see it? [20:44:14] from the page you linked, Missions and Programs -> ASTP Subject Files https://history2.nasa.gov/PDFs/Finding%20aids/Programs/apollo_soyuz_test_project_subject_files_finding_aid.pdf [20:44:27] ah right [20:44:31] there's an Apollo 9 LVOT in the one you linked [20:44:33] I had this file open already, too :D [20:45:44] and also a Saturn Launch Vehicle Systems Handbook for Apollo 5 [20:45:49] I'm only seeing the Apollo 9 SCOT [20:47:11] oh yes, you're right, it's a SCOT, I can't read [20:47:39] LM-4 Systems Handbook [20:48:57] so uhh [20:49:18] patience that it becomes all public or starting to send emails? :D [20:49:26] emails! [20:49:46] that Document Management System of theirs exists on archive.org, but is not currently accessible [20:49:53] so I guess it was/will be public? [20:50:04] could be either haha [20:50:20] lots of coming soon on the website [20:50:55] I am still very doubtful, but I wonder if HQ is taking over all the history stuff from the centers [20:51:04] naaaah :D [20:51:18] that would be too nice, if that is the reason for the JSC History Collection trouble [20:52:40] from all I know NASA works exactly like the German Aerospace Center and Airbus Defence and Space, with each facility doing their own thing and distrusting all the others [20:53:04] that seems way too collaborative [20:53:05] yeah that is definitely true of NASA [20:57:02] it wasn't quite so bad at the German Aerospace Center, it's structured a bit differently. Each center has a bunch of institutes. There is only so much in-fighting you can do as a single institute. [20:57:21] but doing their own thing, for sure [20:58:17] I guess I will pick out my highlight PDF and send them an email, also asking what the general plan for all this is [20:58:27] because a bunch of this seems to be a recent development [20:59:48] sounds like a good plan to me :D [21:05:53] night! [14:38:59] hey [15:00:19] n7275, was there anything else you wanted to check in the ASTP branch? I forgot all about it while I was flying Apollo 12 in the last 2 weeks. You had approved the PR, but then asked for a scenario update, which I did. But new pushes dismiss old approvals of course. [15:08:34] Ahh, whoops. I forgot to re-approve [15:08:54] that should be all set now indy91 [15:09:25] great, thanks! [15:10:13] we have a lot of open PRs again [15:10:57] I am not feeling too bad about it this time [15:11:13] 3 of them are RTCC ones from me, very mergable [15:11:28] 3 of them are from Ryan, also easy to merge once he had the time to finish testing [15:11:34] so 6 of them are easy to get rid of [15:11:44] I have to get back to the failure simulation one at some point... [15:13:06] there's an older failure simulation one from someone else. I wonder if there is anything we should cherry-pick to a newer PR so we can just close it out [15:13:44] I would have closed that older PR when mine gets merged. I had a good look at it and took some ideas from it. [15:13:59] okay cool [15:14:13] my PR was almost ready and then Alex picked too many holes in it :D [15:31:10] what are the chances that Alex appears so I can ask him about a specific commit from 3 years ago [15:58:54] non-zero, but probably not high [16:19:10] I am consulting Alex from the past right now [16:19:12] IRC log [16:19:22] I think I am finding the answers I was looking for :D [16:22:33] morning! [16:24:17] hi Mike [16:59:17] hey Mike [17:22:13] [19:02:25] alright got the VC views to save/load [17:22:22] and the outside cameras to break :D [17:29:19] that must've been a while ago [17:30:01] 3 years ago [17:31:17] There is a flag keeping track if you are in the VC or not, but it's not set for the external engineering cams. So if you save/load in that view you are completely stuck in it. Technically you are still in the VC, but not really [17:31:25] all a bit confusing, also what the intent of all the code was [20:25:51] Those engineering cam views are probably very old [20:33:01] yeah [20:33:13] they used to 80% of what our CSM VC could do [20:33:23] but now they are a bit broken [21:08:26] night! [14:11:51] good afternoon [16:09:27] hey Nik [16:31:25] what's up? [16:31:58] hey Matt [20:55:02] night! [14:49:10] "The document you are requesting is actually a NASA employee only document. The spreadsheet uploaded to our history site needs to be updated to show only those materials which are publicly available. Thank you for bringing this to our attention. " [14:49:16] of course... [15:07:38] classic NASA [16:14:16] that's annoying