[23:00:56] Meeting started by thymo [23:01:12] Meeting chairs are: indy91, thewonderidiot, rcflyinghokie [23:01:30] Uh oh [23:01:31] .topic LM Heat loads [23:01:46] Current subject: LM Heat loads, (set by Thymo) [23:02:40] Mmkay, time to see if I can figure out how to get into a state where I can bring up P63. (I assume I'll need to give the AGC some landing site coordinates, for one...) [23:03:26] And undocking might help too... [23:05:43] Back to the heat loads, I think you are right thewonderidiot, I can just simply calculate heat load by power consumption [23:05:52] Is there a known power usage in standby? [23:10:08] 0.5A times whatever the battery voltage is [23:10:53] You could do Heat(Voltage()*0.5). [23:11:11] Or whatever name ended up being used for that function. [23:13:45] All the heat loads I am using from the chart are assuming 28VDC [23:13:59] So perfect world 14W [23:14:38] sounds good to me [23:15:55] Ugh side note, the more and more I look at the CM ECS the more I want to rebuild it [23:16:40] I just fixed a conversion error in the direct O2 valve and because of the way the code has the o2 flow calculated, it measures a flow higher than actual [23:16:51] But I guess I will deal with that later [23:17:09] Probably after 8.0 release :P [23:17:21] hehehe [23:17:44] The LM is moving along well though [23:19:03] Now that HEATLOAD is a hydraulic function, it is just added to the system timestep [23:19:26] 9.0 will get some CSM love. We'll need to get the SIM BAY going if I'm right. [23:19:37] Yep [23:19:56] And when we push for 9 I will take what I learned for the LM ECS and start breaking the CSM [23:20:24] spontaneously combusting CMPs! [23:20:26] :P [23:20:30] Haha oh yes [23:20:37] Hey if I initialize the suit I am ok ;) [23:20:55] We had another problem like that when we tried adding the hatches and opening them [23:21:05] LM ECS tanks had NaN for energy [23:21:21] NASSP 11 will probably be SKYLAB. AAP might be with that, otherwise 12.0 [23:21:24] Because they werent initialized [23:21:51] What about 10? [23:21:56] we'll have to try to get Skylark out of Margaret for that ;) [23:22:02] LM Block 2 and EVA [23:22:33] thewonderidiot: Indeed. We've got until 2020. :P [23:22:45] I doubt she'll be around that long, unfortunately [23:23:33] Who knows. She's currently 81. [23:25:51] Well the nice thing about the LM is the stage is set to add EVA [23:26:18] Actually we will probably be adding an EVA button of some sort to disconnect the crew from Lm consumables [23:26:26] And not kill them [23:30:09] Also slap the VC and failures somewhere in there. [23:30:59] One thing in the roadmap is "Make MCC use natual-sounding phrases in messages" which I doubt will happen soon or is really any demand for. [23:31:21] Orbiter 2015 support is basically there. Some slight issues with the pad still. [23:32:25] Systems modelling to enable hardware breakage due to mismanagement is coming along nicely. New systems (and ones being overhauled) are detailed enough that adding support for breakage will be easy. [23:33:12] Random failures are sort of there. The C&W has a chance to fail (to often). Engine failures. [23:33:42] Oh and I did a little on the mismanagement already. Pull the plug on a timer and you've lost the track of time. :P [23:54:41] rcflyinghokie: /* ECS without killing the crew... */ if(!NASSP) {return;} if (suit_isol == suit_disc && cabin_press < 1.76 && PLSS) { goEVA(); } if (occupied && !helmet && suit_press < 1.76) { asphxiate(THIS); return("Do you know who Jim LeBlanc is?\n");} [23:56:31] Hehe [23:57:00] managed to get stuff in his stash again and he doesn't know what it is. [23:59:57] Looks like a merge leftover. git stash drop and done [00:00:07] Anyone remember what the LM docking angle is hardcoded as? [00:00:47] 0? Alex or Nik should know. [00:01:13] I never looked at the docking code. [01:07:41] I still have a "NO ATT" light, but I'll take the rough alignment. Now to bring up the AGS. :D [01:11:03] we haven't implemented AGC->AGS communication yet, so it still needs to be done manually [01:11:12] that should be coming in the short term though [01:11:28] rcflyinghokie is the resident AGS usage expert [01:12:01] flips AGS switch to STBY and waits for the case temp to come up. [01:12:13] Sorry was making dinner [01:12:16] Whats up? [01:12:49] haha nothing really, just blathering [01:13:00] Ah ok saw the chat window flashing [01:13:12] But yeah I got pretty good with the AGS :P [01:13:36] I think I am actually going to start there combining the heater for the AEA and ASA with the operation heat load [01:13:39] Make dinner, I'll wait. I'm bringing up an AGS. (The joys of simulated spaceflight: A pause button [01:13:42] Seems simpler than the IMU [01:14:01] Oh I am all done, cooked ate cleaned and now grabbing another beer :P [01:15:45] Scratch that, a whiskey :) [01:15:57] But yeah I got pretty proficient with the AGS [01:16:05] Especially after someone fixed it :) [01:17:08] lol [01:18:40] Which mission so I know which version of AGS [01:19:30] Got the master alarm all working as it should. Now it needs to be hooked up to the CWEA. [01:19:39] Which I will do another day. Goodnight! [01:19:42] Hmm, be nice if I had the AOF vol. 3 for LM-5. I have the one for LM-11+ [01:20:13] Current subject: General, (set by Thymo) [01:20:14] s/AOF/AOH [01:20:24] Yeah same with us [01:20:30] For the most part it is the same [01:20:48] I used it to supplement the LM 8 systems handbook when I did the LM ECS [01:21:48] And *damn* NASA for taking down the good parts of the NTRS [01:21:56] we kind of worked around that [01:22:29] a lot of the stuff that went missing was cached by archive.org's web scrapers, and the filenames have patterns [01:22:50] I wrote a python script to scrape areas around known good documents and pull down all located archived PDFs [01:23:05] so far there's only been a handful of things that we knew were there that we haven't been able to recover [01:23:49] our main source for new documents right now is the UHCL History Collection https://historycollection.jsc.nasa.gov/default.htm?CFID=3072753&CFTOKEN=13277169 [01:24:10] they got a whole bunch of stuff from MSC [01:24:24] ...speaking of which I'm supposed to request the Saturn V Systems Handbook [01:24:34] Ah yes its January [01:24:57] I'll request it tomorrow morning [01:27:23] But back to the AGS, Cannibal^ the only thing we cannot do with it yet is the PGNS/AGS updates so you will have to update the state vector for the CSM and LM manually [01:27:31] And its a pain for sure [01:27:39] coming soon to a LM near you [01:28:07] I know ECS took over but I am surprised on how fast it became stable [01:28:16] yeah you guys killed it [01:28:30] I never would have guessed it would already be working [01:28:40] I was the master of the config, Nik did the C++, and Alex hammered out the meshes [01:28:58] and I distracted people :P [01:29:01] Hahaha [01:29:25] Well now I need to get heat loads going otherwise the LM will be too cold [01:29:45] I need a minimum of I think 240W in order to run the sublimator [01:30:13] And of course without heating of the glycol, the suit temp would always be cold [01:58:42] Actually, I have an interesting question: Is landing via PGNCS possible without touching the AGS? (I know it's unrealistic, but I'm not really concerned with that: a PGNCS -> AGS link failure would've meant an increased likelihood that mission became a non-landing one.) [02:01:32] You mean if the AGS failed would it be allowed? [02:06:17] PGNS and AGS operational was a requirement for DOI and PDI for Apollo 11 [02:06:22] No, I know NASA would say "Enjoy the view, take pictures, jettison the LM, come back." What I'm asking is: If I started P63, could I make it to touchdown without getting aborted due to lack of an AGS [02:06:56] AGS is required through high gate [02:07:21] All the way through low gate [02:07:57] After hitting low gate if you lost AGS you would still be go I believe [02:08:52] But anywhere through PDI ignition and low gate if your AGS was INOP you would be told to abort I think [02:09:48] As in "Nice try, Please perform: do your job, monkey." hardware limitation? or "We monkeys here in Houston would prefer that you have AGS operational." policy limitation? [02:10:25] Policy, its in the mission rules verbatim [02:10:45] Loss of AGS proir to high gate is abort [02:10:52] after high gate its a continue [02:11:00] I guess it was safer to continue than abort that low [02:12:02] Yeah low gate looks like it is in late P64 [02:12:10] "Coffin corner": There are parts of the flight envelope where humans shouldn't be but you can still fly. [02:12:16] I guess you have manual control at low gate [02:13:20] So if I understand properly: nothing stops me from initiating the braking phase and landing without an operating AGS. [02:13:47] Other than a meat-space policy, that is. [02:17:36] Ohh yes [02:17:48] Well a few months ago we didnt even have the AGS working [02:17:54] Did all the landings with PGNS [02:18:28] And even in apollo 11, the PGNS was used the whole way, the AGS only used for reference, until Neil gimbal locked the PGNS when docking :P [02:19:12] I think I thought you were asking about the real mission rules and not the physical limitations haha [02:22:58] Oh, no I've seen a mision plan, I was more concerned with the physical ability. (You can operate a car without a muffler. If you're willing to suffer the wrath of the local government if they catch you, that is./You can land an LM without AGS. If you're willing to be grounded for life by NASA Administration, that is.) [02:24:09] s/a muffler/wearing a seatbelt [02:25:15] Better put, since the AGS is a safety feature. [02:25:25] Yeah pretty much [02:25:45] And since there is telemetry on WAY more than the crew can see, it isnt something they could hide [02:26:32] Got it. :D [02:26:51] However, you could land with just the PGNS no problem [02:27:03] Though it is cool to compare the two going down [02:27:23] I have done many AGS aborts now and its pretty crazy how close the PGNS and AGS are [02:30:18] ... Actually, I'm surprised PAMFD doesn't have a SV->AGS feature (Or does it?) [03:31:06] Sorry for the delay [03:31:15] Use RTCC MFD [03:31:25] It has an AGS SV option [03:31:31] Its not an uplink though [03:31:37] You have to type them in manually [03:48:20] Ah. [04:33:39] hah! [04:33:51] wait is this a transcription error [04:34:49] no [04:34:49] wow [04:35:23] so, you know how you know they never used the output counter EMSD? [04:35:34] every single program we have has it defined with the wrong address [04:37:50] EMSD is, according to the schematics, ND-1021042, and the Symbolic Listing Information, at address 052 [04:37:54] sorry, 056 [04:39:07] Colossus 237 and 249, Sunburst 37 and 120, Retread, Comanche 55, Aurora, and even SuperJob all have it defined as 055 [04:40:18] that is pretty bad [04:43:18] you would think that SuperJob at least would have it right since it was written at Raytheon [04:56:54] What was it supposed to do? [04:57:54] talk to the EMS in the CM, or a "LM monitor" in the LM [04:59:51] specifically to provide the EMS with velocity info [05:00:14] So it didn't suck so much? >:) [05:01:17] lol, I guess [05:16:10] reading ND-1021042, I'd *kill* for board shots of the logic trays. (And as someone who knows some of engineering, "when your change order proposals are able to be bound in a 400 page book, you have a big project. When those change orders are for a single device, you have a *massive* project.) I wonder if GM still has schematics. [05:19:25] yeah, I'm on the same page [05:20:01] the logic modules were all sort of uniformly laid out, with 2 boards each and 60 chips per board [05:20:16] and the schematics get us as far as knowing which gates were on which chip, and which pins were connected to which [05:21:22] I'm working on a database of interconnections right now, with the intent of trying to get physical access to one of the unpotted AGCs so I can beep out the things we don't know (main connector pinout, internal connector pinouts, core memory module pins, some other unlabeled pins, etc.) [05:22:04] Two books. (Good lord. Now I understand the reservations Wally Schirra had about Apollo 7) [05:22:12] lol [05:22:37] and Don's copy of ND-1021042 is even missing a good 100+ pages out of the middle, and doesn't have the supplement [05:23:35] (I would also kill for that supplement) [05:24:00] and proper schematics for all of the B tray modules.... [05:24:12] it's all gotta be out there somewhere, ready to be found [05:35:44] still... I'm really happy that Don had ND-1021042 [05:36:13] I had spent a solid couple of years searching all over the place for schematics of the B tray modules and DSKY [05:36:39] and then one morning just woke up to an email from Don saying he had dug this thing out of a box and if we would like to have it scanned [05:36:40] lol [06:11:31] Wait, B-tray is mostly core rope and core memory, right? So it'd be sense amplifiers, hand-woven core rope and core planes. ND1021042, vol 1 has B7 (oscillator module) at page 283 on the PDF. Betwen Volume 1 and Volume 2 there's a couple tidbits: Two erasable core modules with half of the erasable memory each, 64x32 core planes, stacked 16 deep. (It logically follows that each plane is actually [06:11:31] 32x32 cores, to give 1024 words per module, totaling the 2048 words of erasable. You're only missing B8, since drivers are just power supplies and sense amps are amplifiers and ADCs. I'll wager there was a lot of hand-tuning based on the on-hand component's performance particulars. [06:13:42] Crap, it's snowing again. I gotta go see if my washing machine's drain has frozen up. :/ (The joys of home ownership. yay...) [06:16:26] hopefully it hasn't! [06:17:28] there's only one erasable memory module that has all 2048 words (module B12), with planes of 64x32 [06:19:54] the modules are B1-B6 (fixed memory), B7 (oscillator), B8 (analog alarms), B9-B10 (erasable drivers), B11 (current switch), B12 (erasable memory), B13-B14 (sense amps), B15 (strand select), and B16-B17 (rope drivers) [06:20:11] ND-1021042 has relatively complete schematics of all of them [06:20:34] but sometimes it is lazy about labeling pins [06:21:51] all of the resistors they used for tuning are marked as such, although unfortunately they rarely give ranges for them [06:22:35] https://docs.google.com/spreadsheets/d/1iuJ0ruYgXJvbM6b3NhaP7wdop9FBo5l0OB51hDs4LMc/edit?usp=sharing [06:22:54] that's the spreadsheet I've been capturing pinout information in for all of the B tray modules... in varying levels of organization [06:24:26] it's complete enough to build to, at least [06:24:44] ND-1021042, I mean [06:25:52] the biggest pain in my ass so far are the cores used in the ropes and in the erasable switches... but I think I have some now that should work [09:19:53] . [11:51:47] Hey [11:55:05] hey [11:59:45] any progress with the CWEA? [12:00:21] The master alarm is now connected to the CWEA. Still looking if I should keep this implementation or go even more detailed. Right now I have a single bool MasterFF that is the flip-flop of the MA. In the LM it was triple redundant, I could simulate it down to every relay. [12:00:40] http://www.ibiblio.org/apollo/Documents/HSI-208818.pdf [12:00:42] Page 110 [12:01:48] so, my guideline has been, simulate every relay only if a functional difference could arise from that [12:02:25] in the SECS there are a lot of double relays, so two relays which both get energized under certain conditions [12:02:54] but apart from a relay failure, there would never be a functional difference if we are only simulating one relay there [12:03:07] so in those cases I never implemented the 2 relays, just one [12:03:37] If the first MasterFF would fail relay K1 wouldn't get energised. That would still give you a master alarm but no telemetry. [12:03:45] ah [12:03:49] so that is a functional difference then [12:04:09] looks similar for relays K5-7 [12:04:46] it's not a terribly large number of relays in the CWEA, so I think you could do all of them [12:05:20] I'll do that. [12:05:27] you could take one shortcut [12:05:35] just simulate on flip flop, but 3 relays [12:05:38] one* [12:06:07] and then the if-else for the relay driver just sets all 3 relays as true or false [12:07:10] yeah, I think that's the best compromise [12:07:48] I think my initial assumption was wrong. Originally I though that the master alarm would go off if a C/W Power Caution is triggered. [12:09:39] almost looks like the opposite is the case [12:10:02] hmm [12:11:46] if those relays K5-7 are not energized, so the CWEA has no power [12:12:31] then the 3000 Hz tone can't happen [12:12:36] hmm [12:12:58] the wiring is not so easy [12:13:02] Relays K5-7 even pull the MA breaker to ground on a power failure. [12:13:20] Follow the wire right of the tone generator. [12:14:00] they pull it to ground if the relays are not energized [12:14:06] Yes [12:14:40] so what does that mean [12:14:58] tone generator constantly on? [12:15:21] Hmm now that you mention it. [12:15:30] I think so. Tone generator on, MA lights off. [12:15:49] in the case that the master alarm has power, but the CWEA not [12:15:55] And no way to turn it off without pulling the breaker. [12:25:38] so why do the checklists often have you cycle the CWEA breaker? [12:26:43] I think that resets all the signals through the CWEA reset bus. [12:27:21] ah, I see [14:03:02] Good morning [14:04:37] hey Ryan [14:05:53] So I am worried the changes to direct O2 may cause issue down the road until the CM ECS is reworked [14:06:13] The way the actual direct O2 is plumbed and the way the sim treats it are quite different [14:06:55] But I dont think it will have any awful effects, just will set off the o2 flow light every time [14:07:27] which might be correct [14:08:01] Could be in space, but when on the pad there is a checklist item to adjust for 0.6-0.8 LBH [14:08:44] However since the direct o2 flow is added to the other flows, it measures an o2 flow of 4-5 LBH instead of 0.6 or so [14:09:01] hmm, I see [14:09:18] Because the sensor just adds all flows where the actual ecs system has the direct o2 downstream and it is not an additive function from what I can see [14:09:44] But again, nothing detrimental, just a little annoying for now [14:10:45] It can be played with of course but unfortunately changing the flows is not as simple a fix as I thought to remedy the CM haha [14:11:03] But it should be safe to keep for now [14:12:57] So Mike and I came up with a crude heat load for the LGC in standby [14:13:25] 0.5A at 28VDC would use 14W of power and thus 14W of heat? [14:14:32] sure [14:16:26] Now I other then some power data, the CDU and PSA heat are still a little enigmatic [14:16:30] *Now other [14:17:29] The PSA gets power from both IMU breakers and the LGC breaker [14:18:13] I have a breakdown of the PSA supplies but no amperages [14:18:44] Looks like the national water management agency is closing all the flood barriers. [14:18:51] Quite a heavy storm today. [14:18:57] Where at? [14:19:09] Every one in the country. [14:19:23] Otherwise we flood. [14:19:37] Yikes [14:20:09] yeah, we have quite the storm here as well [14:21:03] https://en.wikipedia.org/wiki/Delta_Works [14:21:33] They are even livestreaming the closure of https://nl.wikipedia.org/wiki/Maeslantkering on the news. :P [14:22:24] The engineering of such structures is fascinating [14:22:49] You guys stay safe out there [14:22:57] Its just 17 and sunny here haha [14:23:01] 17F [14:23:20] so, -8C or so [14:23:54] They lowered the criteria for closing them. Not because of safety but because they think it's exciting to test under actual conditions. :p [14:24:06] Usually 3.1m now 2.6m. [14:24:43] Well the nice thing is they can sell it as a safety thing [14:25:11] Right [14:26:00] The last time they had to close the barriers was in 2007. [14:34:22] hey [14:35:14] Morning [14:36:44] So thinking it over, the increased heat from the LGC breaker when compared with the power on the breaker itself is likely heat generated by the PSA, PTA and/or CDU [14:38:34] yeah [14:38:36] hey Alex [14:39:19] Interesting the IMU OPR has a 200W power load but a heat load of only 78.6W [14:40:08] maybe that's where the increase heat in the e.g. GASTA is coming from [14:41:46] Could be [14:42:31] I think we are just going to have to get it close haha [14:49:01] uhh, can you explain this PR? [14:49:06] about the Direct O2 [14:49:12] what did you change? [14:52:16] Looks like a checklist change [14:52:21] Checklists [14:52:29] yeah, but to what [14:52:41] One notch instead of wide open [14:53:26] I see [14:53:32] For the "adjust for 0.6-0.8 lbh" [14:53:36] is that different from the proper launch config though? [14:53:44] Nope [14:53:52] or is there not "the correct" position for the Direct O2 [14:54:05] For that step its adjust to a certain flow [14:55:26] all good then [14:56:00] The actual verbiage is "DIRECT O2 vlv - open (CCW) (>2 in H2O on SUIT/CAB dP ind) (O2 flow - 0.6-0.8 lb/hr) [14:57:04] Also all the checklists had "LPM ingress" not "LMP ingress", minor spelling change [14:58:01] lunar pilot module [14:58:22] that's the module in which the lunar pilot is stored until he is needed for LM activation [14:58:38] Haha I'll go back and change it... :P [14:58:59] I also don't believe in the SECRA [14:59:02] Hahaha [14:59:14] pretty sure it was a typo as well [14:59:15] Yeah I am not a believer either [14:59:23] I couldnt find anything for that when I looked [14:59:48] And considering every other acronym in that diagram is in the glossary... [14:59:54] Brb [15:15:24] So i am wondering if its prudent to start by taking the heat loads in those systems attached to the lgc and imu breakers, using the power consumption for them, and talking the remainder and divvying it up between the other heat sinks [15:18:29] Then figuring out a clean way to have the imu case vent to the radiator when in standby/on the heater, and give excess heat to glycol when operational [15:22:40] I guess the important part is that the combined heat load is about right [15:23:17] Yeah that will give proper results [15:24:28] With everything running the glycol should be about 40-45 degrees at the accumulator [15:25:04] But 80-100 at its hottest point is possible [15:25:40] The boiler in the real LM can handle like 6000W or so [15:27:00] Of course that's including heat from the pumps and the cooling heat exchanger and such not just systems [15:29:39] When fanning the pumps add Q dont they? [15:33:19] our pump class does not add any energy [15:34:26] What about indirectly via pressure increase [15:35:01] well yeah, that of course it does [15:38:44] Ok thats what i meant haha [15:40:11] cya [16:52:33] Crap [16:52:51] I accidentally went to 100x time acceleration instead of 10x. [16:52:54] CMC Restart. [16:55:03] Had to do a V36 to get the DSKY back. [16:55:26] Parity and TC Trap. [16:55:48] aka RIP erasable memory [16:58:14] State vector is a mess [16:59:59] Both state vectors come up with complete garbage in P21. [17:04:10] IMU still agrees with GDC [17:10:53] Welcome back [17:11:45] My CMC is bust. Accidentally went to 100x acceleration and got a restart. [17:12:11] Both state vectors are complete garbage and I'm pretty sure there's more things that are corrupted. [17:12:27] Where can I fetch the telemetry client again? [17:14:15] good question [17:14:44] I thought it was in the old sourceforge CVS [17:14:59] but you can't seem to access that directly anymore [17:15:26] We really need to give it its own repo [17:15:31] yep [17:15:41] Best of all. ibiblio is down again. [17:17:00] Wasn't there some kind of recovery pad load that could be uplinked? [17:17:05] indeed [17:17:11] we have them for a few missions [17:17:23] and we should create them for the other missions that had that [17:17:39] Apollo 8 didn't have the backup erasable load in the checklists yet [17:17:49] we have Apollo 14-16 [17:18:01] oh, not uplinked [17:18:07] it was in the checklist [17:18:11] That's the one [17:18:20] for manual entry into the DSKY [17:21:00] it's still on Github [17:21:10] in older versions [17:22:45] https://github.com/dseagrav/NASSP/tree/1b507b8a9f0bfb9a319ea789aea76a15589595c5 [17:22:57] this was one commit before it got removed [17:25:21] so, I was cleaning up some LEM code and through that I found that we have a fairly complicated sound file handling for the lunar landing [17:25:25] with many files for Apollo 11 [17:25:41] only problem is, it was connected to the old lunar landing autopilot [17:26:49] so I am not quite sure what to do with all of that [17:27:23] it accessed time since PDI, altitude and a few other things [17:27:31] and then played the right sounds accordingly [17:27:59] I think we need a whole new concept for historical audio anyway [17:31:28] morning! [17:32:28] hey [17:33:23] is the ACA checked at all in P63 or P64? [17:34:44] Hey Mike [17:34:54] yes, you can switch to Att Hold [17:35:13] hmm [17:35:32] but I assume its position is only checked after that [17:35:34] that's normal procedure even [17:35:41] to check manual control before you need it [17:36:13] thewonderidiot: Bit 1 in the restart monitor "E Memory or F Memory Parity Fail". Is there any way to know in what memory that occurred? [17:36:36] bit 2 is set if it was for erasable [17:36:49] Hmm [17:37:07] I don't simulate erasable parity alarms, since there isn't a real way for us to get them [17:37:13] I got a fixed parity fail together with a TC TRAP. [17:39:15] And my REDOCTR got reset [17:40:03] would going from 0 to 42 counts and back be a computational burden right now in yaAGC? [17:40:13] depends on how often it's done, I guess [17:40:31] I haven't connected my joystick for my attempts so far [17:40:32] I was just trying to think of counters that would stop in P66 [17:40:49] so I've been using the keyboard, which is always full deflection of the ACA [17:41:13] I mean it doesn't matter right now since yaAGC doesn't count anything [17:42:18] by the way, I found a silly error in every AGC program we have last night [17:42:23] or most of them, at least [17:42:56] every one that has an erasable definition for EMSD, the unused EMS counter, has it defined at the wrong address [17:43:01] even SuperJob has it wrong [17:43:17] pretty good indicator that they *really* never tried to use it lol [17:43:26] wow [17:43:40] is that an output counter? [17:43:43] yep [17:44:08] so it would probably have been used for reentry and rendezvous, very interesting [17:44:27] originally meant to send velocity information to the EMS in the CM, and test data to a "LEM Monitor" (counter name LEMONM) in the LM [17:44:34] but it wasn't used in either spacecraft [17:44:42] LEMONM at least has the correct address defined [17:44:51] thewonderidiot: Can I force yaAGC to not check parity while still using a parity rope? [17:44:58] I don't think so [17:45:43] LEMONM is probably something that detects if a LEM is docked or not? [17:45:45] you can switch off all of the alarms [17:46:04] something like that -- nothing I have about it goes into much detail [17:46:28] pretty sure the flag for it that gets used is also called like that [17:47:08] no, actually not [17:47:13] just "a LM monitor function" is what Symbolic Listing Info says [17:47:16] where have I read LEMON before... [17:47:30] "Cell assigned for LM monitoring purposes" [17:47:48] So simply setting CheckParity to 0 won't work? [17:48:25] ah, that looks like it would work [17:48:32] maybe I've just seen it in the AGC code, it's just under THRUST after all [17:48:41] yeah [17:52:42] what does ND-1021042 say [17:53:58] just vague references to the "LEM monitoring system" [17:58:07] P65 did not have any program alarms [17:58:20] so I wonder if any manual inputs I am doing are causing a high load [18:00:40] or maybe it's the other way around [18:00:46] the load isn't high enough in P63 [18:00:50] P64 seemed about right [18:06:07] hmmmm [18:06:15] that is a good point, it could be either way [18:06:32] are you generating RADARUPTs in P63? [18:28:53] er, I guess you're already generating them whenever the LGC requests them [18:51:01] yeah, from the LR [18:58:19] well, once we get all of the counters going, we'll have to log them and see how the load changes :D [19:00:52] I'll try a 15/2 run later [19:01:03] but I have a feeling it won't cause any program alarms in P63 or P66 [19:03:34] which is an interesting situation [19:03:41] yeah [19:03:47] In P64 you have a bit more to do generally than in P63 [19:04:04] so the alarms in P63 prepared them that they could repeat [19:04:20] getting the first alarm in P64 would be annoying [19:07:22] although not so critical anymore [19:07:35] you can just not do LPD commands, if you think that is what is causing the alarms [19:07:39] and switch to P66 early [19:44:03] hey Swatch [19:47:36] ok, now I have time to do test landings [19:51:35] :D [19:57:54] with 15/2, no alarms in P63 [19:58:50] got one in P64 though [20:00:53] none in P66 [20:01:06] the COMP ACTY light was on constantly at the end of P63 [20:01:12] but not in P66 [20:01:23] so at least P66 has a slightly lighter load than P63 [20:04:17] interesting [20:04:57] I'll try 22/3 next [20:05:19] sounds good [20:05:20] I reverted some stuff, ExtraDelay to :8 was what we did? [20:05:25] yep [20:05:27] ok [20:07:47] so I bet I will now get alarms in P63 [20:08:01] not early enough to be like Apollo 11 [20:09:27] but at least alarms in the right programs [20:09:56] and the other thing I want to try is using a joystick in P66. Maybe something in the LGC needs more processing at full deflection [20:10:01] although I doubt it [20:10:30] not anything with the counters, but in the fly-by-wire calculations [20:14:16] one other interesting observation [20:14:28] I am pretty sure I got some short freezes in P63 on this previous run [20:14:32] but no alarm [20:14:59] DSKY was not as responsive as it should have been [20:17:43] 22/3, got the P63 alarms back [20:22:44] yep, this run gave me alarms in P63 and P64, but not P66 [20:23:13] COMP ACTY light was on basically constantly in P66, so it was close [20:28:38] oooo [20:28:42] that sounds good :D [20:29:00] 13.6% [20:29:07] that's awesome [20:29:52] we've pretty much independently confirmed that the additional load was not 15% as widely reported, heh [20:31:36] I still think we need more load for P63 and P64 [20:32:04] the order of DSKY inputs actually matters, I noticed this time [20:32:13] I was usually pressing RSET to clear the alarm [20:32:24] and then key release, to go back from the 16 68 [20:32:52] but if I press KEY REL first and then RSET the DSKY doesn't react at first and the chance for an alarm is pretty good [20:33:29] maybe I shouldn't press RSET much at all [20:35:25] hmm [20:36:11] I need to fully reread this section in Don's book [20:37:11] the 1971 Apollo 11 simulation has some timed V16 N68, but no V05 N09 [20:37:24] yeah I was just looking at the 1969 one, same there [20:37:42] so I guess it wasn't trying to exactly mimic the Apollo 11 inputs [20:39:11] how much of an impact would a 05 09 have? [20:39:18] or even the PROG light? [20:39:56] well, every keystroke is a job, and I assume drawing 05 09 is [20:40:07] I'm pretty sure the both launch a job. So a coreset should be available. [20:40:18] PROG light I don't think would have any load [20:40:48] Non additional since that is called from within a job. [20:41:09] V05 will take a job. I had to do TC ENDOFJOB when I wrote my own verb. [20:41:24] Verb 16 is also bad [20:42:02] 16 is quite a heavy load compared to the rest. [20:42:31] yep [20:42:35] That just pushes itself to the background and keeps coming back. [20:43:08] how fast do you respond to radar requests? [20:43:11] what exactly is the logic there? [20:43:41] On a side note. I'm about to go into my first LOI on Apollo 8. :) [20:44:52] I've got a very nice alignment on my IMU. It can point to stars spot on. [20:45:49] haha, awesome [20:46:14] nice :D [20:46:15] radar requests for data from the LGC? [20:46:18] yep [20:46:23] next timestep [20:46:28] so every 1/60 seconds [20:46:38] so radarupt happens next timestep? [20:46:46] yep [20:46:53] hmm [20:47:16] so that's ~16-17ms later [20:47:35] those requests took 80+ milliseconds in reality [20:47:58] so that could potentially be a source of higher than expected load, depending on how the software is behaving [20:48:59] how hard would it be to try to reduce that to 12 or 13 timesteps later? [20:49:15] er [20:49:26] bad math [20:49:29] 5 timesteps later [20:49:53] I could add a timer, wouldn't be so bad [20:50:09] so it would start counting the time when the requests is there [20:50:27] yeah that sounds like it should work [20:50:28] and answer the request (plus resetting it, plus RADARUPT) after X seconds [20:50:31] yup [20:50:55] I would implement that identically in LR and RR I guess [20:50:59] yeah [20:51:07] although it doesn't really matter for what we're doing right now [20:51:08] ok [20:51:11] yeah [20:51:20] and will be fixed properly with the counters, since most of that time is SHINC/SHANC time [20:51:23] so I would use 80ms there [20:51:47] yep [20:51:56] according to Exegesis at least [20:52:13] it says LRHJOB and LRVJOB both sleep for about 80ms while waiting for data [20:53:02] ...it might be wrong though [20:53:18] R-700 shows 80ms for the radar gate, with radarupt showing up much later [20:53:24] hmm [20:53:29] so maybe it's 80ms max? [20:53:48] that's just the time for the radar itself to take a reading, possibly [20:54:01] not accounting for counting the result into the AGC [20:54:50] my old hardware sim screenshot that I posted to the LM thread forever ago shows 110ms [20:54:59] so let's start with that [20:55:06] with 80ms [20:55:12] with 110 [20:55:17] oh, ok [20:55:29] I'll do more detailed timing analysis tonight [20:55:32] I wonder if P66 can process more data, due to lower load [20:55:43] yeah, that could be it [20:55:58] so it can actually work with the flood of data [20:56:03] yeah, I'll implement this [20:56:08] :D [21:00:35] is there any danger that the LGC changes the output bits during the downtime? [21:01:24] what's the order on those? it probably will set the 3 RadarA/B/C bits first and then RadarActivity bit [21:01:47] yeah, that order is right [21:01:58] I doubt there will be any chance of it doing that [21:02:01] I think the ICD warns against it [21:02:13] would that actually happen in the AGC electronics? [21:02:20] LGC* [21:02:27] processing of those bits [21:02:28] that's actually my current direction of study for the counter update [21:02:28] lol [21:02:46] let's see what the Systems Handbook says [21:02:48] bit 3 is particularly dangerous, I think it would switch which radar it was talking to without thinking about it [21:02:54] yep [21:02:58] that's why I am asking [21:03:27] what request signal are the LR and RR actually getting? [21:03:43] http://klabs.org/history/lm-icd/06_radars_sh-48.pdf [21:03:46] that's the ICD [21:04:02] there is a lot of signals and I haven't gone carefully through all of them yet [21:04:42] radar gate select decoder [21:05:13] so the RR and LR are actually never seeing those bits [21:05:25] but it gets decoded before the LGC output interface [21:06:32] looking at the schematics, as far as the AGC hardware is concerned, you can write to those bits at any time and they will take effect immediately [21:06:54] yes, those 3 output bits get decoded to 6 possible signals [21:07:13] LR range, LR velocity X/Y/Z, RR range, RR range rate [21:07:27] going even as far as switching from clocking LR SYNC to RR SYNC or vice-versa [21:07:42] I will write test code for this so we can observe it in the hardware sim [21:07:53] ND-1021042 has a nice table for that [21:08:32] I've also been curious about what happens if you start a radar reading when your 3 output bits map to one of the unused combinations [21:08:45] like 0 0 0 isn't used, I think, so what happens if the LGC requests that? [21:09:01] nothing [21:09:32] RR or LR get a request signal, but not the other required signal that chooses which data is returned [21:10:07] so just 0's get counted in? [21:10:11] so this radar gate select decoder has nothing on the output [21:10:18] gotcha [21:10:32] it might be 0s [21:10:42] probably depends on the RR and LR electronics [21:10:58] yeah [21:11:07] so I'll just use 0 for yaAGC [21:11:10] the shift register will get a readout signal [21:16:46] technically this decoder is the job of yaAGC, right? [21:16:59] because the RR and LR should never see the output bits directly [21:17:01] yep [21:18:21] we'll have to work out the implementation specifics, but I was envisioning something like yaAGC choosing from one of six variables to SHINC/SHANC into RNRAD, based on the output bits [21:18:57] hi Ryans 1 and 2 [21:25:28] not a very elegant implementation, but this will work [21:26:21] should I try 14/2 again? [21:26:26] sure [21:28:04] I'm having an issue with LOI. According to MCC my perilune should end up at 60. But P30 gives me -51.9. Something's wrong here. [21:28:07] using 0.11 seconds delay [21:28:15] yeah, P30 is wrong [21:28:30] it just adds the DV vector to the state vector at ignition [21:28:53] it doesn't simulate the burn as a multi minute, non-impulsive burn [21:29:13] for long burns this will usually result in a pretty bad HA, HP estimate [21:29:53] as an example, Apollo 11 LOI-1 [21:30:01] Okay. Was getting kind of worried. [21:30:08] 072:51:24 McCandless: The values which you would see on Noun 42 prior to LOI burn are HA, plus 431.3; HP, minus 128.2. [21:30:54] the HP value will probably depend a lot on the altitude at ignition [21:32:44] this question identifies you as a LOI first timer :D [21:33:11] hehehe [21:33:35] Oh definitely. [21:34:17] ok first attempt with a 0.11 seconds delay of the LR response [21:37:59] So if I read this correctly they didn't do ullage on LOI-1? [21:38:22] no, I don't think you ever do it with a SPS propellant tank this full [21:38:27] there is a chart for this [21:39:04] https://history.nasa.gov/afj/ap15fj/csmgc/5-08.gif [21:43:50] So anything right of the 2-jet ullage time does not require ullage? [21:44:58] yep [21:45:25] thewonderidiot, had no program alarm in P66 [21:45:43] that is good [21:45:46] what about P63/P64? [21:45:50] loads [21:45:55] oh wow [21:46:00] okay, so this was a good change [21:46:13] I can't 100% remember what this 14/2 combination did before [21:46:41] (01:06:39 PM) indy91: 14/2 report [21:46:42] (01:06:53 PM) indy91: I had to annoy the LGC quite a bit to get an alarm in P63 [21:46:43] (01:07:20 PM) indy91: only very late in P63 and when I did the KEY REL on the 16 68 [21:46:43] (01:07:29 PM) indy91: P64 was basically as bad as usual [21:46:44] (01:07:48 PM) indy91: alarms every so often and one after I looked at 05 09 [21:46:44] (01:08:06 PM) indy91: one alarm in P66, a few seconds before landing [21:46:45] (01:08:09 PM) indy91: 1202 [21:46:51] ah [21:46:56] so it might have been luck [21:47:06] the COMP ACTY light was on constantly in P66 [21:47:54] I think I completely understand the duty cycle diagram now [21:48:08] https://www.doneyles.com/LM/dutycycle@146x450.jpg [21:48:10] well if you had loads of P63/P64, you should be able to back off with the RR CDU [21:48:40] loads as in, in P63 I got alarms if I had the 16 68 up for 10+ seconds [21:48:53] after V57 of course [21:48:55] haha [21:49:21] do you have the diagram up? [21:49:24] from back to front [21:49:42] the last peak is switching from P66 to P68. [21:49:53] it doesn't need to do any real time calculations anymore, so it can do 100% [21:50:13] then, just before, is on the surface in P66, with the LR having no good data anymore [21:50:16] that helps with the load [21:50:20] I could see that in the sim [21:50:32] COMP ACTY light was not on all the time [21:50:39] then before that is P66 [21:50:46] which is only slightly lower than P63 [21:50:58] the plateau is all P64 I think [21:51:17] yeah, that sounds right [21:51:23] and then the long one before P64 is P63 [21:51:31] with only small increases when the LR does stuff [21:51:41] or starts doing calculations [21:52:08] the actual calculations aren't so complex for the LGC [21:52:18] for the data incorporation I mean [21:52:42] calculating range and velocities from the raw data is worse [21:53:09] and the LGC does that whenever it has a data good signal, before V57 [21:53:36] I almost had the diagram figured out before, just P66 after the landing got be confused [21:54:17] any idea what the bottom line might be? [21:54:46] er, there are two lines graphed below that [21:55:01] not really [21:55:10] maybe components of the complete load? [21:56:17] possibly yeah [21:56:49] interesting that it doesn't stop in P68 [21:56:55] at the very end [21:57:02] so maybe RR? [21:58:31] I'll do some more testing tomorrow [21:58:34] good night [21:58:50] night! [22:38:51] About to pass behind the moon! [22:39:18] 10 minutes until ignition [22:40:24] There goes the earth. Catch you on the flip side. [22:40:37] LOS [23:01:16] Lunar orbit :D [23:01:24] \o/ [23:03:13] Damn. She sure is beautiful. [23:04:37] Just over 50 years later Apollo 8 is in lunar orbit again. Albeit in a virtual simulation. [23:07:20] just over 49, we haven't hit the 50 yet :D [23:08:45] Oh right. December 24 wasn't in 2018. D'oh [23:18:30] And AOS with Earth. :D [23:19:38] I can see India, Madagascar, Africa. Awesome! [23:19:52] Australia, Indonesia [23:20:15] Middle-East area [23:20:24] Antartica [23:31:10] man I really need to keep flying 8 [23:55:21] Arrgh, why does D3D9client keep jacking up cleartype!? [00:01:52] That's not D3D9 [00:02:27] In the Launchpad go to Extra>Debugging options>Performance options [00:02:47] Uncheck "Disable font smoothing while Orbiter is running" [00:08:38] ...I always forget that. [00:23:30] Took me a while to figure out the first time too. [00:23:43] I don't think it's a very good default. [00:27:09] Oh, I think I know why telemetryclient was ditched: You get CMC, three elements in EPS, and uplink. For the CM. And the connection is a bit flaky. [00:30:02] It'd be nice to have, but right now it's in need of a sln for vs17. She doesn't compile at least for me. [01:02:42] That's good to know. Thanks. [01:27:23] .in 10h AGC scaling [01:30:08] thewonderidiot: How accurate do transcriptions need to be? [01:30:32] what do you mean? [01:32:07] I saw a misspelling of CMC as CGC. That turned out to be correctly transcribed. However I also spotted REQUESTWITH in the scans and it was transcribed as REQUEST WITH. [01:32:37] Just making sure. :P [01:32:43] we weren't as pedantic with whitespace as with non-whitespace, but it is always nice to correct [01:34:47] It's in the assembly information of Artemis072. http://www.ibiblio.org/apollo/ScansForConversion/Artemis072/0023-P1090638.jpg [01:34:49] 00115 alarm [01:34:59] lol [01:35:12] I take it that means it's in almost every program [01:35:31] Uhh [01:35:34] Let's check [01:35:41] Colossus 249 [01:35:43] hm, 249 and later [01:35:48] not 237? [01:36:10] no, it's correct there [01:36:47] It's also wrong in 249 [01:36:59] yeah, 249 and later [01:45:39] I'll make a PR tomorrow (or you can do it if it's more convenient). Time for some sleep now. [01:45:41] Goodnight! [02:08:08] night! [03:36:56] Question: The AOT detents are X (in N70/N71) 00xSS for the mark routines, (P51/P52) and the detents are themselves numbered LF=1, F=2, RF=3 RR=4, R=5, and LR=6? I understand the AOT (At least I think I do) has the spiral as the X-mark and the three lines as the Y-mark, yet I'm getting massive angular differences during P51. (+04163 as an example) [03:43:07] Cannibal^: yeah, there's something unknown wrong with the AOT right now [03:43:31] we have the equation for the spiral so it is probably not that... so it might be some sort of field-of-view problem or something [03:43:53] Niklas, Alex, and Ryan probably know more about it [03:44:05] I'm just repeating what I've collected listening in [03:44:29] So how do I establish a good REFSMMAT so I can try for a landing? [03:44:53] Alex had some sort of cheaty way to get good AOT marks [03:45:02] let me see if it's in this computer's logs... [03:46:59] hmm, yeah, not sure [03:48:01] ah, it was a cheat for P57 [03:56:10] anyways, I know very little about using the AGC operationally or actually flying missions, heh [03:56:19] unfortunately everybody that does is asleep [03:56:32] grumbles about getting program alarm 220 after using RTCC MFD's REFSMMAT upload [04:39:32] well, here's a more detailed view of what the radar interface looks like [04:39:43] this is the AGC reading Y velocity from the landing radar [04:39:47] https://i.imgur.com/yaGDG5h.png [04:41:31] .tell indy91 here's a more detailed waveform of the radar interface: https://i.imgur.com/yaGDG5h.png [04:42:06] .tell indy91 it looks like the reading starts at the same time as the first TIME5 increment after CH13 bit 4 is set, so 110ms is the absolute worst case [04:46:16] hm, I might have a wiring error -- it's generating gate pulses for 90ms instead of 80ms [04:57:49] this may be an initial conditions problem [05:06:59] this isn't cleared by GOJAM though, so the very first radar reading is going to be bugged [05:07:04] which is weird [05:12:59] .tell indy91 actually that trace is wrong, it's gating for 90ms instead of 80ms. turns out there's an initial conditions problem with the radar circuitry. the real number should be ~90-100ms, not 110ms [05:13:55] .tell indy91 I don't really see any way in which the very first radar pulse would be correct, though, at least one RADARUPT has to happen for the circuits to be initialized correctly to generate an 80ms gate pulse train [06:16:55] cages IMU [06:34:19] FL V06 N05; R1: +04413. [06:34:50] :/ 44 degrees? THEY WERE IN THE SAME DETENT! [06:48:16] re-enables the optics debug string [06:56:21] Hmm, OpticsShaft isn't changing. Where's my precision hammer? [07:23:28] Hmm, maybe it *is* turning. But reversing directions 0->1>2->1 goes 0->1->2->3. That's not right... [11:01:40] 6.43 degrees! In an P57 align using detent 3. That... ...isn't terrible. A lot better'n 44.13 degrees [11:16:43] :/ 32.64 degrees on detent 4. Twice. Exact same accuracy. That ain't gonna fly. [12:16:07] Hi [12:18:47] Hey Niklas [12:19:10] hey [12:20:16] ok :D [12:20:53] I wanna see if I can get yaTelemetry to work. No luck so far. [12:22:36] with which AGC version? [12:22:43] Any [12:22:56] It doesn't want to connect to yaAGC at all. [12:23:37] And the VirtualAGC manager is not working, can't find the binaries and it doesn't play nice with my KDE color scheme. [12:23:44] works for me, when I run the Virtual AGC through the GUI [12:24:05] Wait [12:24:26] Got it [12:24:31] Wrong directory. [12:24:40] Had to go into temp/lVirtualAGC/bin [12:25:03] yeah, the directory structure is a bit weird [12:25:18] I'm looking into the LM timestep [12:25:23] http://www.ibiblio.org/mscorbit/mscforum/index.php?topic=491.msg22750#msg22750 [12:25:40] there is this weird 4 timestep delay in the code [12:26:03] when you load the simulation, it doesn't do any systems timesteps for 4 timesteps [12:26:20] I hope I can remove and fix that [12:27:20] I noticed that. Had to implement a fix for the mission timer so it wouldn't lose the time due to no bus voltage. [12:27:58] oh, where is that fix? [12:28:32] the mission and event timer in the LM also didn't have a systems timestep, so they didn't draw any power [12:28:44] Just the bus voltage I think. It had a custom one but I changed it to the default min dc voltage. [12:28:55] They didn't? [12:29:24] I phrased that wrong [12:29:26] Not even in any other code? I'm certain I put in power calculations based on the amount of ELDs lit. [12:29:30] the systems timestep wasn't called [12:29:38] Oh [12:30:16] I moved all the systems timestep in a separate function, as it works in the CSM [12:31:14] and now I'll try to find the issue that made the 4 timestep delay necessary [12:32:33] hmm [12:32:46] and your systems timesteps for the timers don't check if the timers are even powered [12:32:54] so they always draw power [12:33:33] Let me look at the code [12:33:48] MissionTimer::SystemTimestep [12:34:36] Haha oops [12:34:49] I was very much a PanelSDK noob when I wrote that. [12:34:50] it should be something like [12:34:53] if (IsPowered()) [12:34:54] DCPower.DrawPower(11.2); [12:34:54] if (IsDisplayPowered()) [12:34:55] DrawPower(7.0 * 7.0 * 0.022); [12:34:57] Yeah [12:35:09] I'll fix that [12:35:17] Thanks [12:38:21] CM event timer would only need the DC power, right? [12:38:32] Yes [12:38:42] That has a mechanical display. [12:38:44] so I'll do this there: [12:38:45] if (IsPowered()) [12:38:46] { [12:38:46] if (Running) [12:38:47] DCPower.DrawPower(5.0); [12:38:47] else [12:38:49] DCPower.DrawPower(1.0); [12:38:51] } [12:39:12] Looks good. [12:40:14] ok, commited [12:41:11] in that forum post dseagrav said that the LM is using the PanelSDK wrong somehow [12:41:25] I bet that is also what is causing the occasional bad frame rate [12:41:30] that gets fixed by pressing any key [12:43:05] Also it looks like Cannibal^ had some issues with the AOT this morning. https://vanbeersweb.nl/irclogs/%23nassp/2018-01-02-23:00_Untitled-meeting.log [12:43:22] oh, I am dumb. You actually can browse through the old CVS commits still [12:43:23] http://nassp.cvs.sourceforge.net/ [12:43:28] the telemetry client is also there [12:44:10] I'm beating on it right now. the best accuracy I got with a surface align (P57) was 6.43 degres. [12:44:12] Cannibal let me know that it doesn't compile in VS2017. Needs a new solution. [12:45:47] Cannibal^, which programs are you having problems with? Any of the alignment programs? [12:46:17] Alex did some changes to the panel, I haven't done an LM IMU alignment since [12:46:38] I'll quickly try a P52 [12:47:44] Right now, I'm trying to figure out why the AOT is giving 30+ degree angle differences. I got a halfway decent one out of one detent, but the rest were atrocious. [12:49:47] It's *almost* like the AOT view angles are wrong, and that's what I'm playing with right now. (Trying to figure out how to get the camera vector spat into the DebugString [12:55:11] no, we spent a lot of time to get those view angles right. I am 99% sure they are correct in principle [12:55:32] I just did a P52, +00003 angle difference [12:55:50] of course that's the easiest case [12:56:03] only the front detent, already a good alignment etc. [12:56:41] which mission are you flying right now, Apollo 11? [12:56:59] also, what monitor resolution do you have? [12:57:28] maybe our panels don't support greater than 1080p or so [12:58:40] Apollo 11, and I'm running windowed 1920x1080 [12:58:45] ok [12:59:21] and which program is giving you the 30+ degrees? P51? [13:00:09] P57, landed at tranquility. [13:00:32] ok, so I guess you had a good alignment before landing [13:00:46] And it'd be option 4. [13:01:08] we have some P57 accuracy issues with the spiral, but they should not be more than 1° [13:02:26] I think I'll roll back to an older save and try an align again in orbit. I didn't have the best. [13:04:04] option 4, so a new alignment to the landing site [13:04:11] Yep. [13:04:14] which alignment technique? [13:04:20] 2 stars? [13:05:26] I think the good one was one body and gravity. [13:05:59] That'd be method 3. [13:06:48] yeah [13:07:14] I guess you used the AOT counter on the AOT panel? [13:07:21] to determine the spiral angle [13:07:41] that one is also fairly new and I haven't tried it. Before we had to use a debug line for this [13:07:49] so potentially something is wrong with the counter [13:09:26] nah, looks good [13:21:59] oh, the LM RCS failure flags are buggy [13:22:12] when you reload a scenario they all show the red flag, so the failure state [13:35:12] fixed [13:39:24] In orbit. V37E00E. PRO. V22E 00512E PRO. mark-X, mark Y, V21 00515E PRO. mark-x, mark-y. PRO. double digit star angles. [13:40:15] Sorry, V21 rather than V22 [13:40:24] that doesn't look quite complete [13:40:33] what after the V37E00E? [13:41:06] P52? [13:41:19] Yep, sorry, [13:42:24] and then in P52, option 3? [13:43:27] Arrgh, damnit. That was in p51, sorry. [13:43:42] ok [13:43:50] and you used the rear detent [13:43:52] CL [13:44:15] Yes. [13:46:51] I'll try a P51 with the rear detent, in the scenario that gave me a good P52 [13:47:09] did you coarse align the IMU at the beginning of P51? [13:47:13] not that it really matters [13:49:58] Good morning [13:52:43] grumbles. [13:53:19] FL V06 N05, R1: 04774 [13:54:30] For a P52? [13:54:37] hey Ryan [13:54:40] I just did one with +00002 [13:54:47] For a P51. [13:54:47] so it must be a procedural error [13:55:08] I even used the same stars as you [13:55:09] 12 and 15 [13:56:13] I actually don't use the X/Y mark feature much. I usually mark when the star is exactly centered [13:56:37] but you should be able to do it fairly off center, as long as the star is on the X or Y line [13:56:50] Oh AOT [13:57:10] yep [13:57:16] I got pretty good with the X/Y marking on those [13:58:04] Q to mark on the X-line, Y to mark on the Y-line, right? [13:58:11] on QWERTY at least [13:58:34] Yep [13:59:22] I could try it in the exact same scenario as you [14:00:11] maybe something is wrong with the scenario, somehow [14:00:27] the detent angles are padloaded [14:00:39] so if your erasable memory is broken in some way, it would give fairly random results [14:00:57] Also you are using the correct detent? [14:01:01] if you give me the scenario you are using I can check all those things [14:01:05] The one the LGC gave you [14:01:09] P51 [14:01:14] LGC doesn't give you anything [14:01:17] Ah yeah [14:01:22] I keep thinking 52 [14:01:41] he was using the rear detent (CL) and I just did a good P51 with that [14:02:17] Also when you get a chance, would you help me decipher what exactly they are doing with the IMU init? [14:02:23] line 474 or so in lemsystems [14:02:49] They have some heater stuff in there and also a bunch of commented out stuff [14:03:57] yeah, that's weird [14:06:33] can you defined those parameters in the config? [14:06:36] define* [14:08:19] I am not sure about area [14:08:32] But you can define a mass and isolation [14:08:48] I need to look into area and see how it is used [14:08:56] I assume surface area for a radiator though [14:08:59] yep [14:09:03] for thermal effects [14:09:21] In which case it's probably a definable parameter in the config [14:09:37] Why it is defined in here I don't know [14:09:41] radiative energy [14:09:58] I don't think it's done this way anywhere in the CSM [14:15:15] oh, I see [14:15:23] when a radiator is created from the config file [14:15:24] new_one->Area = (1.0 / 4.0 * volume); [14:15:46] and volume is defined in the config [14:16:08] Interesting [14:16:14] parameter size in h_Radiator [14:16:47] and size is used for the calculation in the radiator timestep [14:16:49] A = 1/4*V? [14:17:23] interesting conversion :D [14:17:42] Yeah [14:18:15] Back in a few [14:42:21] we need to make a decision regarding mission audio [14:49:06] Ah yes. You mentioned something along the lines of a complex simulation with Apollo 11? [14:49:36] What kind of mission audio? [14:50:41] Apollo 11, during the lunar landing [14:50:46] so I was cleaning up some LM code [14:50:59] and wanted to remove a few unused variables in the LEMComputer code [14:51:11] and those are/were used to play audio at certain times [14:51:41] how it worked was that the old lunar landing autopilot was updating a bunch of variables [14:51:51] altitude, time since PDI and a few more [14:52:04] and depending on these numbers, certain sound files were played [14:52:12] altitude callouts for example [14:52:18] obviously all of this is broken right now [14:52:22] Right [14:52:32] so we can do two things [14:52:51] remove all the audio files and drop support for PDI specific audio [14:53:10] or add certain events, maybe like TLI, EOI in the CSM [14:53:16] and fix the audio [14:53:52] It would be cool to have some mission audio [14:53:53] connecting it to the Virtual AGC would be pretty tricky [14:54:04] I mean the launches have it [14:54:14] But on timers and events [14:54:23] yeah, although that works slightly different [14:54:37] there is a special class and file called SoundEvents [14:54:55] Why couldn't landing audio be based on events outside if the vagc? [14:55:06] it could be [14:55:13] we just need to reimplement that [14:55:28] and calculate the numbers that the sound events needs [14:55:56] Seems more straightforward than a vagc link or so [14:56:08] yeah [14:56:10] we need: [14:56:21] simcomputert, MissionTime, mode, timeremaining, timeafterpdi, timetoapproach [14:56:47] I'll have to research what some of these were [14:57:35] Sound\ProjectApollo\Apollo11\sound.csv [14:57:44] that's the file that defines what sound is played when [14:59:57] hmm [15:00:05] I'm not sure this was ever a completed feature [15:00:54] I know some missions have landing audio on contact, would the idea be to add this sort of thing to all missions? [15:00:58] these variables are not used in the old lemautoland.cpp [15:01:48] I mean personally i am ok with only major event audio but im sure more immersive audio would be appreciated [15:01:49] we are talking about close to 100 sound files for Apollo 11, all played at specific times or altitudes etc. based on the logic with these time parameters I posted [15:02:35] oh, now I understand. It was functional before [15:03:26] making it work again, without the old landing autopilot and without a direct connection to the Virtual AGC will be quite difficult [15:04:38] I mean this is one of those finishing touch things, again im not against it but it just seems unnecessary at this juncture [15:06:44] Don's landing simulation do it really hacky, they just look at the DSKY program, verb and noun to start an event like e.g. astronaut input [15:06:49] simulations* [15:07:13] we could do the same theoretically [15:07:27] check for P63, and the right verbs and nouns to start the PDI event [15:07:46] that would only have to look at the DSKY, not the Virtual AGC [15:08:17] I suppose altitudes could use the dsky as well [15:09:09] well, that could also just access the altitude from the Orbiter API [15:09:32] But unless you are landing on autopilot, there is a lot that could become out of sync [15:09:59] ah, the sound event code checks for altitude by itself [15:10:16] just have to do a few fixes for Orbiter2016 compatibility [15:11:48] I'll familiarize myself with this sound event code [15:13:58] When i get time im going to start playing with that imu heat [15:14:36] But i have to play catchup for work first haha [15:14:55] ok, so the decision is, keep the sound file support, if possible [15:21:09] ACTION: Should PDI specific audio support be maintained? [15:21:19] AGREED: keep the sound file support, if possible [15:26:06] indy91: I did LOI-1 yesterday. How should I go about this P52? I read on the forums that they did it differently in the mission and you yourself ran into some issues with the ORDEAL. [15:26:31] The flightplan dictates option 3 but should I do an option 2 instead? [15:26:47] oh, the ORDEAL issue was ages ago [15:26:58] these issues have long been solved [15:27:11] you keep the LOI-2 REFSMMAT for the entire stay in lunar orbit [15:27:20] that is from before MCC-4 to after TEI [15:28:36] Haha oh wow. [15:28:53] I think I had used a option 1 REFSMMAT for my first ever Apollo 8 LOI-1 [15:29:04] before we were able to calculate the LOI-2 REFSMMAT [15:29:07] I just loaded a scenario that I thought was where I left off. But I found myself in space with the moon nowhere to be seen. [15:29:15] Turns out I was passing around the dark side. [15:29:40] yep, it's really dark [15:30:26] the option 1 REFSMMAT for LOI-1 is basically the opposite direction of a nominal alignment for orbit [15:30:34] that's why I had ORDEAL issues [15:30:42] it was rotating the FDAI in the wrong direction [15:31:10] the REFSMMAT that MCC calculated for you should prevent that [15:31:27] and it should give you nearly 180°, 0°, 0° angles for LOI-2 [15:32:17] no wait, in RPY [15:32:22] 0°, 180°, 0° [15:32:30] that's the ideal [15:32:40] the actual Apollo 8 mission was 5° off in pitch [15:32:57] but I've seen our RTCC/MCC doing a better job even [15:36:19] I didn't get the LOI-2 pad yet. LOI-1 had a pitch of 160° [15:37:14] yeah, that's probably good [15:38:52] LOI-1 was very good. Got 168.6-60.3. Very close to the PAD values. [15:39:27] yeah, that's pretty good [15:40:41] morming [15:40:41] One fun thing to note. The HGA still has a signal even when behind the moon. [15:41:33] Hey Alex [15:42:48] hey [15:42:52] Thymo, that should not happen [15:43:12] I implemented the HGA in such a way that should prevent this [15:43:20] maybe I didn't do it for the omni antennas [15:43:24] Indeed. I'm blind. There is Earth. [15:43:25] False alarm. [15:43:54] haha [15:44:04] HGA and OMNIs should all loose signal [15:46:08] Do you plan on doing HGA auto tracking some time? [15:48:24] the problem I had with that in the past was the 3-axis movement [15:49:16] the CSM HGA doesn't just have a pitch and yaw axis, but it has a third axis, that only gets used by auto tracking, when the HGA comes too close to the singularity at 0/0 angles [15:49:50] basically the antenna would have to be rotated by 180° very quickly at those angles, to keep auto tracking [15:50:02] so the third axis prevents that, somehow [15:50:21] I don't have enough information about when the auto tracking uses that third axis [15:50:47] and the geometry calculations for the 3 axis movement made my head hurt [15:51:03] I'll figure it out eventually, I hope [15:51:18] Haha. No rush. [15:52:47] Our IMU and GDC perform too well. I have scrubbed all GDC alignments so far because I simply do no see any difference when I switch between IMU/GDC. [15:52:56] yep [15:53:20] the ASA is not quite as good [15:54:00] there is some drift due to the AEA calculating the attitude from increments [15:54:18] but the GDC and IMU are basically perfect [15:54:28] the BMAGs I should say [15:56:35] at least the ASA is simple to realign, 400+3 [15:56:53] as long as PGNS is up of course [15:57:08] yeah [15:58:41] We should add random drift. [16:00:41] One thing that would be nice in RTCC MFD is a way to calculate a back-up star alignment for AGS. (Send star set, Reticle angles to MCC and then it gives the correct attitude for the 400+5 [17:33:49] morning! [17:38:18] hey Mike [17:39:12] what's up? [17:44:38] nothing much messing around with the LM right now [17:55:22] Hey [17:55:26] hmm [17:56:00] would there be any interest in yaAGC simulating the radar interface bug? it'd have to be probability-based since yaAGC operates on too long of a timescale [18:01:33] I haven't been keeping up with how you two have been torturing the AGC the last couple of days. I'd vote for as accurate as possible. indy91's call. [18:04:06] Not knowing what the vote is for id agree with the accurate as possible ;) [18:04:43] lol [18:04:43] (09:55:26 AM) thewonderidiot: hmm [18:04:44] (09:55:59 AM) thewonderidiot: would there be any interest in yaAGC simulating the radar interface bug? it'd have to be probability-based since yaAGC operates on too long of a timescale [18:06:58] Is there a way to make a flag so that it only simulates if failures are enabled? [18:07:31] sure [18:08:39] If that's the case then absolutely [18:13:16] it only really messes with 5, 9, and 10, I guess [18:13:23] since 11 had a hardware workaround in place [18:13:31] and 12+ have C13STALL [18:14:13] hey Mike [18:14:21] o/ [18:16:02] well, what do we have to do if the radar interface bug were to be implemented? [18:16:22] all I know about it is the panic it caused before Apollo 10 [18:16:41] nothing on your side, just a few lines of code in yaAGC that will rarely garble the data you give me [18:17:26] is that the same thing you talked about yesterday? That makes the initial data coming from the radar bad? [18:17:36] no, that's different and was never fixed [18:17:52] That was resolved by a recycle of the breaker right [18:17:53] it only affects the very first reading of any radar, a single RADARUPT fixes it [18:18:08] I think that was different [18:18:11] would VHF Ranging also have that issue? [18:18:19] yep [18:18:28] Because the VHF Ranging had a bad first mark [18:18:35] every time, every mission basically [18:18:55] the first use after AGC poweron has something like a 15/16 chance of generating the wrong amount of gating pulses [18:19:04] I've read that in a few debriefings [18:19:07] :D [18:19:11] action was rejecting the first mark [18:19:14] that is good to hear, lol [18:19:19] and from then it never exceeded the threshold [18:19:29] well now we know why! [18:19:39] yeah [18:19:57] it wasn't much of a mystery [18:20:10] just something they noticed [18:20:19] they probably heard about it in a CMC lesson [18:20:28] yeah, not a big deal if it's only the first one every time [18:21:09] it's really not even a 15/16 chance of it being wrong on poweron... probably something like 15/16 AGCs will get it wrong the same way every poweron, more likely [18:21:34] yeah, that sounds like the description in the debriefing [18:25:09] and when I was transcribing the schematics I chose the wrong initial conditions for every single gate in the pulse counter, hhe [18:25:10] *heh [18:25:25] so that's why I was seeing 110ms instead of the expected 90-100 [18:25:46] rcflyinghokie, an update about LM timestep. I have a few issues to solve in the LM before I can properly move some timesteps to PreStep. So it will take a while. [18:26:44] 90 ms or 110ms, that's not much of a difference on our side [18:26:55] true [18:27:17] Ok no problem. Should i hold off on adding more system heat until then? [18:27:38] nah, you can continue as you did [18:27:55] it doesn't add any additional work for me [18:28:13] it's possible that you will suddenly get a more stable LM ECS [18:28:18] but we will have to see [18:28:41] Ok then. I might play with splitting up the LGC and IMU heat into the PSA and CDU etc [18:28:49] dseagrav had to implement a cheaty way to make LM initialization work, after loading scenarios [18:29:01] But i will start with the systems we have modeled [18:29:07] it has to do with a PanelSDK issue in the LM [18:29:17] Hmm undoing cheats can be annoying [18:29:19] probably the same problem that is causing the occasional bad frame rate [18:29:29] which gets fixed by pressing any key or switch [18:29:35] you are getting those, too, right? [18:29:35] Right [18:29:38] Yep [18:30:07] Every time id press RSET or hit a switch before doing anything [18:30:29] Otherwise it was just choppy and unstable [18:30:51] yep [18:33:10] When using the ECS test scn i always gad the habit of cycling the glycol pumps to clear it haha [18:34:23] yeah, I hope fixing that will also make the initialization hack unnecessary [18:34:58] and then I can move timesteps without causing any new problems [20:38:24] Finally getting around to watching an episode of Enterprise. Meant to do that for a while. [21:33:26] Welcome back [21:35:44] Thank ya [21:36:02] On and off attempting to be productive [21:36:31] Really wishing i could be adding LM heat now that i have an idea how to do the IMU [21:47:25] But it will probably be the weekend [21:49:00] I've watched an episode of ST: Enterprise this evening. Hadn't taken the time since we talked about it. Also been downloading new kernel due to Meltdown. [21:49:38] Nice [21:49:58] I miss college breaks sometimes haha [21:50:12] me too [21:50:38] Meeting ended by thewonderidiot, total meeting length 168582 seconds