[21:49:54] NASSP Logging has been started by thymo [21:51:53] Got a new IP so it'll take a while for the website to get back up while DNS propagates. [21:52:58] that's annoying :( [21:59:56] I still need to finish it up, the modem is currently resting on the vacuum. I'll need to pull the old stuff out and mount the new stuff in. [22:00:48] Considering it's 12AM and my alarm is set at 0630 for work tomorrow I think this is perfectly find until I feel like looking at that shit again. Goodnight. [22:06:57] goodnight! [11:45:44] .loggingstatus [12:27:14] morning [12:27:56] hey [12:28:15] investigating two issues right now [12:28:33] the first one is probably just a procedural thing [12:28:53] S-IC/S-II staging is not enabled until TB2+17 [12:29:20] there is a S-IC function (not EDS) to shut down all the S-IC engines if two adjacent outboard engines are failed [12:29:28] that is enabled at TB1+134.4 [12:29:54] so you could get a S-IC stage that is completely dead, just waiting for the time when it is allowed to do staging [12:30:17] and all this provided you have already disabled two engines out auto abort [12:30:28] I think you are supposed to do a manual abort in this case [12:30:42] it made orbit though [12:30:44] buuuuut [12:30:47] with the second issue [12:30:53] 14km state vector error [12:31:23] ouch [12:31:48] is it because it went out of powered flight mode for a bit? [12:32:05] hmm [12:32:36] in LVDC terms it should never be out of powered flight mode [12:33:03] that is switched at TB5+100 only [12:33:09] right [12:33:55] so, with two engines failed the Saturn will have drifted a bit out of the launch plane [12:34:03] so I wonder if it can't properly deal with that [12:34:13] some sort of crossrange LV IMU issue [12:34:40] the error is also in altitude though [12:34:46] DSKY showed 110NM [12:34:55] instead of the usual ~103NM [12:35:21] weird [12:35:36] and I have 3 FC lights [12:35:42] I think low temps [12:35:44] my SV after 1 engine out insertion was good, about 1.1km [12:36:59] yeah, pretty normal [12:38:43] at the time of the engine failures the error was already 1000m though [12:38:59] in S1C flight? [12:39:18] yeah, I had two at 100 seconds [12:39:32] hmm that seems high for that time [12:39:45] it jumped to 420 meters at liftoff [12:39:52] found a memo about aborts at that time [12:40:07] arguing it will be ok to not abort in all cases [12:40:41] wonder what made it jump to 420m [12:41:13] yeah [12:41:57] Sensed Acceleration: 9.831316 [12:41:57] Gravity Acceleration: 9.817841 [12:42:03] that is just before liftoff [12:42:13] the sensed acceleration is varying a bit which is of course weird [12:42:22] but then at TB1+0 [12:42:35] Sensed Acceleration: 3.825110 [12:42:35] Gravity Acceleration: 9.817867 [12:42:46] one timestep later [12:42:46] Sensed Acceleration: 10.465550 [12:42:47] Gravity Acceleration: 9.817918 [12:42:52] that is probably normal [12:42:59] but the one timestep with 3.8 m/s² [12:43:00] not normal [12:43:16] right [12:43:23] the addforce? [12:43:55] should be part of the sensed acceleration [12:44:08] LV IMU is run in pre step [12:44:14] other accelerometers in post step [12:44:32] I'll check if that gives better results [12:44:33] my last launch had SV of 24.895719 at TB1+0 [12:44:43] but of course LV IMU is only part of it [12:44:47] LVDC navigation the other [12:46:37] hmm [12:46:44] there is one weird line of code [12:47:23] it is weird that the sensed acceleration is varying in TB0 [12:47:51] that potentially is the prestep /poststep thing [12:47:59] its almost like if its oscillating like that very old EMS problem we had when we 1st started using NASSP in Orbiter 2016 [12:48:18] yes [12:48:33] what was the fix for that? [12:48:33] that was using the wrong timestep length [12:48:37] oh [12:48:52] because post step changed in Orbiter 2016 [12:49:01] so I had to modify the accelerometer classes [12:51:26] I just made a change that has the SV error at 0 post liftoff [12:52:12] not sure if that is really the solution for the whole launch [12:52:21] I really don't know why that was there [12:52:39] every guidance cycle (1.7 seconds) the time since the last cycle is reset [12:52:49] so that navigation is done every 1.7 seconds and a bit [12:52:55] depends on timestep length [12:53:07] but there was a line of code modifying that number at TB1 start [12:53:29] not sure what that is there for [12:53:37] I'll do a whole launch [12:54:48] oh really [12:56:31] my understanding was that it causes a navigation cycle at liftoff or so [12:56:59] but it's been there for a long time [12:57:02] I didn't add it [12:57:49] if you have a super stable frame rate then it's probably no problem [12:57:59] GRR at T-17 seconds [12:58:04] 1.7 seconds navigation cycle [12:58:07] so that is synced [12:58:42] 57 meters error at halfway into orbit [12:59:12] and that could easily be something with pre/post step [12:59:49] so maybe deleting that one line of code will be a massive improvement of the LVDC navigation [13:00:00] oh man [13:00:47] 83 meters at 8 minutes [13:00:52] was this an issue in Orbiter 2010? [13:00:53] although it doesn't go up steadily [13:00:53] wow [13:01:03] one timestep earlier it was 84 meters [13:01:10] I guess it was but in Orbiter 2010 we all had a million fps [13:01:12] could also be neglecting J3 and J4 [13:01:20] I always had 60 :D [13:01:43] 87 meters error at staging [13:02:17] I'll have to read deep into the old LVDC development thread [13:02:20] maybe I find something on it [13:02:31] what line is it exactly? [13:03:15] sinceLastCycle = 1.7 - simdt; // Rig to pass on fall-in [13:03:19] in the Saturn V LVDC [13:03:24] sinceLastIGM = 1.7 - simdt; // Rig to pass on fall-in [13:03:33] is the equivalent in the Saturn IB LVDC [13:03:42] ok [13:04:20] 112 meters at cutoff [13:04:24] so I comment that to test? [13:04:27] yes [13:04:30] it's not need I think [13:04:32] nice! [13:04:41] is that on par with the real LVDC? [13:04:43] navigation should start at GRR and just continue through liftoff [13:04:50] it's better than Apollo 12 :D [13:04:59] haha [13:05:01] but a bit worse than normal still, I think [13:05:13] Apollo 12 had a bad accelerometer in one axis [13:05:22] at liftoff already I think [13:05:29] so not caused by the lighting strike [13:06:03] jumping back and forth between 125 and 129 meters [13:06:09] and then the cheaty SV update is done [13:07:05] yeah [13:07:11] probably still needed I guess? [13:07:20] hmm [13:07:31] maybe it's good enough now so that we can apply normal mission rules [13:07:47] update based on difference to CMC and ground tracking [13:08:25] might make MCC-2's a little higher but yeah [13:09:33] or not because while the SV might not be zeroed at insertion, the actual trajectory of the S/C is closer to what the TLI pressetings expect [13:11:15] I have been testing CSM/LV sep on the pad so much lately I now sometimes hit it by mistake instead of GDC align on launch scenarios :D [13:11:24] haha, wow [13:11:40] Apollo 11 Saturn V Flight Evaluation Report only has velocity difference numbers [13:11:47] not absolute SV position error [13:12:09] 1.52 m/s vertical, 1.73 m/s crossrange, 0.54 m/s downrange [13:13:36] ah, more numbers [13:13:40] I can reconstruct their error [13:14:00] great [13:14:16] oh wow [13:14:31] it's a bit more than 100 meters :D [13:14:46] interesting [13:15:12] 457 meters in X-axis [13:15:13] btw I guess this is good news for you S1B launch window targeting [13:15:17] which roughly is vertical [13:15:25] -1280 meters downrange [13:15:31] no [13:15:34] crossrange [13:15:38] -239 meters downrange [13:16:03] so kind of similar to the velocities [13:16:10] but this is on platform coordinates [13:16:11] just trying an insertion myself with the change [13:16:51] SV of 20m at 66s [13:17:29] "the LVDC looses 1.7 sec at liftoff" [13:17:35] is what an old forum post says [13:18:04] maybe something was misinterpreted from the documents? [13:18:54] maybe [13:19:59] I don't think we have this issue at TB6 start [13:20:16] it might be a doing a long first powered navigation timestep [13:20:23] but that shouldn't hurt too much [13:20:55] "Found it. I forgot to set sinceLastIGM to 0 in the GRR routine. " [13:21:04] but that is and should be done [13:21:10] but not anything at liftoff [13:22:27] I have a sv of 21m at TB4+10 [13:23:02] my fps has been constant 144 though, external view [13:23:26] at liftoff most timestep dependency should be gone now [13:23:45] and the rest might be pre/post step [13:23:56] so fps should not matter really? [13:23:56] which I could change and potentially fix [13:24:06] fps will matter with the pre/post step [13:24:20] and make it more accurate then Apollo 11 :P [13:24:40] let me look up the Apollo 12 values [13:24:42] for shock [13:24:52] wow 102.7x102.6 [13:25:04] thats the most circular Ive seen [13:25:31] thats off PAMFD [13:25:46] weird thing is V82 shows 103.8x98.9 [13:26:42] LVDC SV accuracy of 25m at insertion [13:27:15] Id definitely not need an LVDC SV update before TLI [13:27:26] CMC isn't perfectly accurate either [13:27:43] starts accelerometers at liftoff [13:28:14] and CMC uses a different altitude reference than PAMFD [13:28:57] how about we remove the cheaty SV update, then make something in MCC tracking that checks the error, then updates it properly if needed [13:29:28] and maybe something similar in RTCC MFD [13:30:02] I've tried and failed in the past to make a better IU SV update [13:30:05] but I can try again [13:30:20] I know quite well how it should work from the EDD [13:31:06] so there weirdly isn't a big error for Apollo 12 [13:31:18] but downrange is 3x as high as on Apollo 11 and that is quite significant [13:31:25] because that is propagating over time [13:31:39] so after TLI the error is very large [13:32:29] 36400, -1100 and 10500 meters error [13:32:30] at TLI [13:35:50] Apollo 11 mission rules only have a IU SV update for accelerometer failures [13:36:04] oh [13:36:08] for later missions there are specific numbers [13:36:22] at GET 56 minutes [13:36:30] downrange error 87801 ft [13:36:38] semi-major axis 3.1NM [13:36:45] crossrange velocity 16 ft/s [13:37:03] Apollo 15 did an update [13:37:03] those seem high [13:37:13] that is from Apollo 14 [13:37:33] so with less then 87801 ft downrange error, they proceeded with no update [13:37:42] yes [13:37:47] Apollo 15 has smaller numbers [13:37:56] I guess that any error can be easily corrected with MCC's [13:37:57] 48575 ft in downrange, 1.6 NM in SMA [13:38:19] I even have the RTCC document converting MSFN, CMC and LVDC SVs for comparison [13:38:52] thats 26 km downrange error [13:39:04] for Apollo 14 [13:39:29] yeah [13:39:41] that's basically just 3 seconds off [13:39:48] so it sounds large [13:39:54] right [13:39:59] but is not super bad [13:40:29] so even with our old SV accuracies of 1-2 km at insertion, we technically could of proceeded with no SV update [13:40:36] ah, I have the RTCC document for another mission rule [13:40:44] TLI go/no-go [13:40:48] just compares LVDC and CMC [13:40:55] those have to be in agreement by some margin [13:41:04] independtly from ground tracking [13:42:43] I have done TLIs without SV update [13:42:56] in Orbiter 2010 that was not as bad [13:44:17] but maybe now it's also good enough :D [13:44:42] the 1-2km accuracy was at insertion though [13:44:54] the numbers I said were for 56 minutes GET [13:45:23] so lots of time for errors to propagate [13:47:43] ah, right [14:03:51] not even in our oldest LVDC version I could find does this line of code make sense to me [14:04:17] it definitely causes a navigation cycle at exactly liftoff [14:04:22] but not sure for what [14:20:33] yeah [14:20:40] just got the 3 FC's come on too [14:21:00] think its caused by the FC RAD temp low [14:21:08] yeah [14:21:33] remember, we had some idle power load to stop that from happening [14:21:53] an idle power load which consumed power over a whole mission and didn't really allow for flying complete missions :D [14:22:26] yeah' [14:25:38] so Ive disabled the SV update, and im running the sim to 56 minutes GET [14:26:24] 56 minutes, must be when Honeysuckle is tracking I guess [14:26:37] or Hawaii [14:30:22] 1.9 km at 56 GET [14:31:37] good enough [14:32:29] hmm at 5+2646 it was 991m [14:33:24] then at 5+2654 (8 seconds later) it jumped to 1879m [14:33:40] and then probably back [14:34:07] what I am not sure about is what really is "now" for the LVDC [14:34:46] LVDC runs in prestep, so the state vector from the Orbiter API is from the end of the previous timestep [14:35:03] hmm [14:35:13] but that wouldn't really explain a 900 meter jump [14:35:25] yeah [14:36:59] it jumped 900m but it did not come back down, 1.9-2km is the new "middle point" and its jumping 100-200 m around that [14:37:13] interesting [14:37:56] any unsual in the lvlog? [14:38:09] accelerations, total velocity etc. [14:38:30] in any case, I'll commit this fix [14:38:38] definitely improves things [14:38:55] ok [14:39:27] https://www.dropbox.com/s/xcdv7h9qmb950ez/lvlog.txt?dl=0 [14:39:51] happens at 5+2654 [14:41:49] yeah, found it [14:43:10] was this during time acceleration? [14:44:26] hmm I had run 10x up to that point, but I think when that was taken I was at 1x [14:50:28] so at GET 2:45 SV is 4670m [14:51:07] oh interesting [14:52:10] a few seconds before it was 5400m, and then a few before that 4500m [14:52:28] si t seems to just jump back and forth 700-900 meters a various times [14:52:55] weird [14:53:17] https://www.dropbox.com/s/ncycqjqi1ale33k/lvlog2.txt?dl=0 [14:53:19] latest [14:53:32] I think Ill try a full TLI now (without the SV update) [15:07:29] one thing we should try to do is cancel the normal mission audio when abort is initiated [15:18:11] Apollo 11 insertion SV 500m~ [15:22:15] which is a bit higher then what we were getting before [15:23:38] but [15:23:56] at GET 56m its at 1.1km which is better then my last one [15:38:26] 3.5 km just before TLI [15:45:51] MCC-2 18 fps [15:46:18] IU SV accuracy is 6.7 km at TLI cutoff [15:48:13] heres the whole log: [15:48:28] https://www.dropbox.com/s/r3tpk3h4whe2yvw/lvlog_A11.txt?dl=0 [16:07:07] MCC-2 18 fps, thats a calculation right after TLI cutoff. To put in prespective, the same calculation in my last Apollo 11 run, was 22 fps [16:21:51] the error seems close to what Apollo 8 had [17:10:17] yeah [17:10:28] you mean the real mission or in NASSP? [17:11:32] real mission [17:11:41] looked it up in the flight evaluation report [17:11:56] ah ok [17:12:01] trying 15 now [17:12:38] but so far looks like it holds up well with no SV update [17:15:49] Apollo 10 was more accurate [17:16:01] maybe 4km error at TLI [17:16:20] maybe the jumping of the SV 500-800 might be more of an issue with SVCompare then the SV itself [17:17:53] possible [17:28:24] although I wouldn't know what [17:29:19] so, did the SV have an error in Orbiter 2010 at liftoff? [17:31:17] yes [17:31:40] my lvlog from my Orbiter2010 installation says so [17:34:52] yeah, the same error applied [17:35:03] but maybe the touchdown point stuff made it more random [17:36:09] 3.8 km error before Apollo 15 TLI [17:43:15] MCC-2 is 6.8 fps [17:44:09] acceptable [17:44:53] Ill check one of my previous post-TLI Apollo 15 scenarios to compare [17:52:17] 5.2 fps on my older scenario [17:53:24] the error seems to vary quite a bit between missions [17:53:37] but lots of missions are about as accurate as we seem to get now [17:53:47] a few a bit lower and a few muuuuch more errir [17:53:50] error [17:55:04] was there anything to be done with the prestep/postep stuff you mentioned? [17:57:07] I have to test [17:57:08] maybe