<awang>
lamont: Does the jaggers code include launch azimuth calculations?
<lamont>
so the code i’m working on now launches at 96.4 degrees to hit the moon which should be broadly correct
<lamont>
PEG doesn’t really do a launch azimuth calculation
<awang>
Does MJ still need the launch-once-and-restart thing?
<lamont>
no
<awang>
Bornholio: Thanks! I'll check that out
<lamont>
there’s the lead angle thing, but i always use zero
<awang>
lamont: Is the lead angle thing the text box that appears next to the launch to target plane button?
<lamont>
if you play around with it you could fine tune it a little bit, but zero should work fine enough for the moon
<lamont>
yeah
<awang>
Hmmm
<awang>
I've always set that to 0
<lamont>
i don’t even know if that code works
<awang>
Sometimes MJ filled it in itself
<lamont>
yeah that just isn’t that important for launches to the moon from the cape
<lamont>
for launches to a specific 90 degree inclination orbit it can make a large difference
<awang>
Inclination difference is too small?
<lamont>
yeah for the moon it doesn’t matter
<lamont>
if you tell PEG to hit the 90 degree inclination orbit that is directly overhead when you launch, then PEG will happily do exactly that, and burn a ton of extra dV to do it.
<awang>
Burn a ton of extra dv in what way?
<lamont>
it wiggles around a lot in order to hit exactly that plane
<awang>
I see
<awang>
So what exactly does the value in that text box mean?
<awang>
And how would I use it to stop PEG from wasting all that dv?
<lamont>
well for the kind of 90 degree inclination launches that we do there’s code from jaggers that allows PEG to rotate the plane around to find a roughly minimum dV solution
<lamont>
if you want to target a space station in a 90 degree inclination orbit and launch-to-plane you’d need to input some value in that textbox that produced a similarly near-minimal ascent when you targetted your space station, but its entirely guess-check-refine
<lamont>
best thing to do would be to setup a 90 degree ascent with the plane free and see what heading PEG suggested first, then tweak the textbox until you got the right leading angle that produced a similar heading
<lamont>
i have no idea what the magnitude, or even sign of that number should be though
<lamont>
(or even if the code in MJ is buggy or not)
<awang>
Remind me to never ever have to rendezvous with anything in a polar orbit then
<awang>
That sounds nightmareish
<lamont>
most likely there’s a number that’ll work pretty good there for most launch, i just don’t know what that number is
<Bornholio>
use standard launch to plane and then haul that number to peg
<awang>
Hmmm
<awang>
Asked MJ to launch into a 150km x 150km orbit
<awang>
Got a 199km x 117km orbit
<lamont>
some kind of terminal wiggles
<Bornholio>
thats good for an old rocket :P Need an RCS circ burn or three :)
<awang>
This was with the vmiss thing you suggested
<awang>
How bad of an idea is lowering the terminal guidance time?
<lamont>
that is likely to be worse instead of better
<lamont>
probably just better to wait until i get the next revision ready
<awang>
I can do that
<Bornholio>
modelling nukes on friday may be inspired to test some peg also. Whats latest good version? Jaggers 3?
<awang>
jaggers3 is the latest tag
<awang>
Use commits after that at your own risk
<lamont>
you might want to hold off for a second
<Bornholio>
New release soon!
<Bornholio>
or soon(tm)!
<lamont>
i’ve almost got launches to the moon repaired again
<lamont>
although i guess the nodeexecutor is probably an unusable mess
<Bornholio>
No Breaking Node Exec! pretty pretty please
<lamont>
yeah i think i’m going to have to back that out again or something
<awang>
Are you seeing the weird headings when launching to plane of the moon too?
<lamont>
that was a “learning experience” and now i understand what more of the yapping in all the PEG papers were all yapping about
<lamont>
no
<awang>
And who cares about node executor, I have Principia :P
<awang>
What was broken about it?
<lamont>
i was trying to make PEG target the orbit after the burn, but was pulling nonsense numbers of of the orbit to feed to PEG and i didn’t understand that’s not really how you do it
<awang>
Aside: Tried a 185km x 185km orbit, got a 151km x 232km orbit
<awang>
engine cut off pretty much right at 185km though
<awang>
Wait, target after the burn?
<awang>
How does that work?
<lamont>
nodes in KSP are infinitely impulsive, so the problem to solve is not to execute the burn, but to target the orbit that the impusive burn creates
<lamont>
so if you’re burning to raise the apoapsis of an orbit you want to look at the desired orbit and target the point where the apoapsis is (3 constraints) and then target the velocity at that point (another 3 contraints) and then execute the burn and let PEG get you onto that orbit
<lamont>
which will magically match the inclination and periapsis and everything else
<lamont>
but my first guess at trying to solve that problem was a hot mess
<Bornholio>
lol, sounds like my attempts at good solutions to supersonic Lox injection calculatiosn for Triton Nukes
<lamont>
so with one guidance law i wound up in a 186.868 x 185.288 orbit with 8530 vac dV left, lets see how a different one works…
Wetmelon has quit [Ping timeout: 183 seconds]
Bornholio is now known as borntosleep
<lamont>
186.961 x 185.333 with 8531 left
TM1978m has joined #RO
<awang>
Almost as if you didn't switch guidance laws at all
<lamont>
yeah think i’m going to go with the simpler one
<awang>
You sure you actually switched? :P
<lamont>
yep its commented out
<lamont>
dropping the terminal guidance from 40s to 20s produced 185.998 x 184.958 with … oh 8522… boiloff, lolz…
<awang>
That seems like a long time for terminal guidance
<awang>
What does terminal guidance do differently from "regular" PEG?
<awang>
Also, what was the difference between the two guidance laws supposed to be if they produce such similar results?
<awang>
Also, my gawd the black arrow had an awful TWR
<lamont>
its basically the same problem in terminal guidance, but the PEG thrust integrals seem more unstable than the atlas-centaur PEG thrust integrals, so the terminal period needs to be longer
<lamont>
although qualitatively its slightly different. atlas-centaur guidance just stopped guidance entirely. the shuttle PEG code still runs guidance but unlocks the terminal position constraint, and freezes the angular velocity of the turn rate to stabilize it
<lamont>
i suspect for the last second or two i should freeze the velocity-to-be-gained vector as well, which might tighten up the terminal conditions even more….
<lamont>
oh and i think the other guidance law might have been better for very long burns around a large central angle
<lamont>
i’m going to leave it commented out, it might turn out to be useful for low-TWR burns to hyperbolic orbits or something…
<awang>
Jeez, guidance is complicated
<awang>
So basically terminal guidance tries to lock down some things to try to ensure guidance still converges?
<lamont>
it always diverges
<lamont>
so you know how mechjeb burns and the blue burn vector on the navball never goes to zero but always wigs out to one direction or the other? that’s a failure in terminal guidance
<lamont>
what MJ should really do is lock guidance for the last few seconds onto the node vector and then calculate how much it needs to burn, and then just burn that much and stop
<lamont>
instead of doing the ‘hunting’ it does sometimes
<lamont>
185.852 x 184.926 with 8529
<lamont>
if i squint my eyes really hard that looks better
Senshi has quit [Quit: Leaving.]
<awang>
Ah, I see
<awang>
That makes sense
<awang>
So the thing that Squad put in that switches stock SAS from follow prograde/maneuver to hold attitude when velocity is low enough is terminal guidance of sorts?
<awang>
Also, tfw ISP is increasing so fast that your current dv left increases during a burn instead
<awang>
Actually
<awang>
I think something's broken
<awang>
...Yep
<awang>
I'm just dumb
<awang>
Forgot I had infinite propellant on from earlier experimentation
aradapilot has quit [Quit: Leaving...]
<lamont>
awang: i just pushed a mchenry1 tag which should have stable ascents — but the node executor is going to be just totally broken
SirKeplan has quit [Ping timeout: 186 seconds]
SirKeplan has joined #RO
ferram4 has quit [Ping timeout: 186 seconds]
Rokker has quit [Quit: Connection closed for inactivity]
Wetmelon has joined #RO
qwertyy__ has joined #RO
qwertyy_ has quit [Ping timeout: 198 seconds]
qwertyy__ has quit [Ping timeout: 186 seconds]
ProjectThoth has quit [Quit: +++out of cheese error+++]
Rokker has joined #RO
Wetmelon has quit [Ping timeout: 183 seconds]
<awang>
lamont: Will I need to change vmiss in-game?
<awang>
Do sun-synchronous orbits HAVE to be retrograde?
<awang>
Couldn't you head southeast to get the same inclination, just orbiting in the opposite direction?
<awang>
I haven't used them, but I seem to remember them being ring-shaped when I moused over them in the parts menu
<BadRocketsCo>
Actualy it may not look too bad
<BadRocketsCo>
The engine fairings from FASA might make it look quite presentable
<BadRocketsCo>
I'm away from home right now so I'll just have to wait until I get home to test it out
<awang>
Now I want to see if I can make the proc parts decoupler ring-shaped
<awang>
And adjustible height
Senshi has joined #RO
<awang>
Hmmm
<awang>
I seem to be getting KSP crashes if MJ tries to autostage when I'm in map mode
<awang>
All I get in Player.log is "Receiving unhandled NULL exception"
<awang>
Followed by "Obtained 9 stack frames"
<awang>
Then no stack frames :(
Hypergolic_Skunk has joined #RO
schnobs has joined #RO
ferram4 has joined #RO
Senshi has quit [Quit: Leaving.]
<awang>
lamont: Get a 196km x 180km orbit when just launching east
<awang>
Target 185km x 185km
<BadRocketsCo>
Getting that accurate of a orbit with anything more powerful than the 1km thruster will be a pain in the behind
<BadRocketsCo>
1kn*
<BadRocketsCo>
I don't think literaly any rocket has ever donr that
<BadRocketsCo>
done*
<awang>
lamont seems to get pretty good results with PEG
<awang>
185.852 x 184.926
<awang>
idk what rocket he's using though
<lamont>
that is dipping into a TLI stage with a TWR of around 0.6
<lamont>
awang: are you correctly using launch-to-plane to hit the moon? because it should be 96-ish heading
<lamont>
and yeah, i suspect the issue with the circular orbit is just how many dV is dropped in a physics frame
<lamont>
i’ve found the correct way to launch to the moon is to: 1. target the moon in the map view. 2. click “launch to plane” 3. make sure the angle is zero 4. click “enable autopilot” and you should then warp around until the plane is over KSC and launch with about 96 heading
<lamont>
and in the map view you could make sure that when you look at the moon’s orbit edge-on that the dot for KSC is directly lined up
<lamont>
and i’ve definitely button mashed the ascent autopilot and wound up with weird results, but none of that is the fault of the PEG code it is just weird MJ behavior
TM1978m has quit [Remote host closed the connection]
BadInternetCo has joined #RO
BadRocketsCo has quit [Ping timeout: 183 seconds]
<awang>
lamont: I'm pretty sure I'm using launch-to-plane correctly?
<awang>
What you just described is exactly how I do it
<awang>
MJ launches at the right time
<lamont>
if you’re getting 120 degree heading though something is wrong
<awang>
Or at least I think it does, I'll need to check after this sim finishes
<awang>
Rendezvous planner says relative inclination is ~3.2 degrees, which I think is about as small as it can get?
<lamont>
rendezvous planner?
<awang>
And I'm running at a pretty low max physics delta-t
<awang>
0.02
<awang>
MJ's rendezvous planner?
<lamont>
you don’t want to use that
<lamont>
PEG doesn’t support launch-to-rendezvous
<awang>
I'm just using it for the relative inclination reading
<awang>
Unless that's wrong?
<lamont>
that would be wrong
<awang>
Oh
<lamont>
launch with a relative inclination of 0
<awang>
The relative inclination never actually reaches 0 though?
<awang>
It oscillates between 3 and 50 degrees
<awang>
Or somewhere around those values
<lamont>
uhm
<lamont>
should be 0.25 or something?
<awang>
Oh
<lamont>
but if you don’t touch the rendezvous button and use 0 then MJ will align the plane as best it can
<awang>
Uh
<awang>
Hmmm
<awang>
Might Principia mess with things? Since it might model the oscillation of the Moon's inclination
<awang>
But yeah, I'm not touching anything that says rendezvous
<awang>
Just launch to plane
<lamont>
hrm
Rokker has quit [Quit: Connection closed for inactivity]
<awang>
egg|work|egg: ^?
<lamont>
if its anything its more likely the J2 of the Earth’s gravitational field that principia models
<lamont>
which will mess up the gravity integrals, but as the burn gets close to zero that should converge
<lamont>
and it shouldn’t be that wonky to start with
<lamont>
you might want to try to launch without principia and see what happens
<lamont>
and make sure you’re on the mchenry1 tag
<awang>
I'll give that a shot
<awang>
I'm on the mchenry1 tag with changes from upstream dev
<lamont>
the jaggers stuff i’ve abandoned and its probably all buggy in one way or another
<lamont>
okay
<lamont>
i’ll give principia a shot, and i’ll start testing with a higher TWR ending stage and see if i can fine tune burn termination a bit
<awang>
I'll see you on the other side :P
aradapilot has quit [Ping timeout: 207 seconds]
BadInternetCo has quit [Ping timeout: 383 seconds]
aradapilot has joined #RO
aradapilot has quit [Ping timeout: 183 seconds]
Calexal has joined #RO
aradapilot has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
Calexal has quit [Ping timeout: 180 seconds]
<awang>
TIL sounding/weather/comms/etc. payloads get refueled by launch clamps
<schnobs>
even lead ballast :)
Senshi has joined #RO
NomalRaptor is now known as UmbralRaptor
<lamont>
awang: 185.004 x 195.842 with 8.7g at SECO in stock
Rokker has joined #RO
schnobs has quit [Ping timeout: 183 seconds]
<lamont>
that is a delta-v error of 3.21 m/s which is 0.0377 seconds at 85 m/s of thrust acceleration
<lamont>
instead of tgo < 0 i’ll check for tgo < TimeWarp.fixedDeltaTime which seems both correct and should back it up exactly one frame
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
Wetmelon has joined #RO
ProjectThoth has joined #RO
<awang>
It never occurred to me how much of an effect such short times at high accelerations can have
<awang>
3.21 m/s also seems strangely small for a 10km difference between apoapsis/periapsis
<awang>
But intuition doesn't work too well with rocket science
Hypergolic_Skunk has quit [Quit: Connection closed for inactivity]
<awang>
Hmmm
<awang>
Does RO or RP-0 delete the accelerometer part or something?
<awang>
It's not showing up in my ModuleManager.ConfigCache