<wb99999999> okay so...lamont are you here?
<wb99999999> found an extreme case that cause slip-ups on PEG
stratochief has joined #RO
TM1978m has joined #RO
<lamont> heh
<lamont> here
<wb99999999> hey
<wb99999999> let me try to recreate it first
<wb99999999> found it yesterday
<lamont> vids are good if you can record it
<wb99999999> ok
<wb99999999> I am going to try to recreate it first
<wb99999999> yes, it happens again
<wb99999999> I called it a slip-up because I am uncertain on weather it was on me or on the code
Raidernick has quit [Read error: Connection reset by peer]
Raidernick has joined #RO
<wb99999999> still need time to figure out recording...
<wb99999999> so what happens is I have a heavy rocket with 1.2 lift off TWR
<wb99999999> and I set it to pitch at .4 deg/s, and no time limit for the pitch program
<wb99999999> naturally the trajectory became a bit too deep and vertical before it hits the desired pitch (and before close loop kicks in)
<wb99999999> but it never hit that pitch, at about 2km/s and 40 deg pitch, the remaining pitch number started to increase
<wb99999999> the close loop guidance never kicks in
<lamont> is “PEG Status: CONVERGED” in green?
<lamont> and does the degree countdown go negative?
<wb99999999> yes and no
<lamont> oh remaining pitch increases because PEG is becoming more and more steep faster than the pitchover is becoming steep
<wb99999999> yes
<wb99999999> but that doesn't help the situation
<lamont> yeah i’ve seen that one
<wb99999999> I am already going too fast and too steep
<lamont> yeah you can’t really get out of that one
<wb99999999> I mean I can just switch to timed pitch program
<wb99999999> but this is one case that might cause someone else trouble
<lamont> it will, but i don’t quite now how to go about fixing it
<lamont> maybe a button to force guidance to PEG right now
<lamont> but that makes it worse like you said
<lamont> and fundamentally PEG doesn’t grok atmospheres at all
<lamont> we could do something like...
<wb99999999> maybe an option to force the PEG to engage when remaining pitch started to increase?
<wb99999999> it never should increase in the first place
<lamont> that paper describes simulating a gravity-turn-to-closed-loop ascent and then shoving it into an optimizer and letting the computer figure out the right way to patch them
<lamont> oh “too deep and *VERTICAL*”
<wb99999999> yes
<wb99999999> so too little pitch
<lamont> gotcha, yeah, then that button would help
<lamont> you can actually edit the pitch rate live
<wb99999999> usually I'll just use timed pitch program
<lamont> and i think the emergency pitch adj is also live
<lamont> so you could change “0.4” to “0.5” and it’ll jerk fairly hard down
<wb99999999> yeah
<wb99999999> but maybe you want some sort of preventative measure built-in
<lamont> tick the AoA limiter before you do that and it should smooth it out
<wb99999999> personally timed program is just fine for me
<lamont> i’ll think about it, but i think if you don’t get the pitch program in the ballpark, then its going to have issues…
<wb99999999> yeah...the situation itself is really on me
<lamont> this is one of the many reasons why I wouldn’t replace the default stock ascent profile with PEG because i don’t want to have to explain all these issues to game players that expect to hit a button and have it work...
<wb99999999> because I built a new vehicle and am not familiar with its trajectory yet
<wb99999999> and got lazy XD
<lamont> yeah, experience and sim runs
<wb99999999> but I think it helps to let you know anyway
<lamont> i mean we’re supposed to be doing booster development runs in RP0 and getting your gravity turn completely wrong is one kind of failure...
<lamont> yeah, i’m thinking about what to do
<lamont> the “right” thing to do is simulate it, but that means extracting the PEG routines into something that can run in a sim and adding travity turn integration and then… etc...
<wb99999999> or maybe just have a tickbox that says "no overshooting pitch program"
<lamont> well its not really overshooting, PEG is just doing PEG stuff, and your pitch program didn’t keep up
<wb99999999> well in that case I overshot
<wb99999999> PEG was fine, but I overshot
<lamont> i think we have different meanings of overshoot
<wb99999999> wait, yeah
<wb99999999> that's not the right verb I was looking for
<lamont> there might be something that could be done to ensure that the delta in degress always decreased, and to kick the pitch rate up if PEG started crawling away
<lamont> clamping so that it doesn’t go the wrong direction would be trivialish
<wb99999999> problem being what if someone intended this to happen, and uses the stage-activation method
<lamont> right
<wb99999999> I can't believe I am still capable of thinking XD
<lamont> i probably wouldn’t patch the peg-after-stage option though, the user told us to only activate after that stage, so we trust them, period, i think.
<lamont> it can also happen with the timed-pitch-to-gravity-turn option though
<wb99999999> I think a easy solution is just to let off a message
<lamont> wish NK was doing vids because he’s a great playtester
<wb99999999> says "your pitch program is busted, do something now" and let players deal with it
<lamont> hah yes
<lamont> i’ve been meaning to patch prevent-unstable-ignitions to throw up a warning like that
<wb99999999> sounds fine to me
<wb99999999> but why does it yaw me so much?
<wb99999999> shouldn't the launch azumith includes the deviation due to Earth's rotation?
<wb99999999> oh I see...
<Starwaster> lamont: So, I had a situation where PEG was converged but status was stuck in unguided gravity turn
<lamont> i’ve patched it (locally) so that won’t happen
<Starwaster> new release past #10?
<Starwaster> oh, LOCALLY
<Starwaster> ok
<lamont> no, its not even committed yet
<Starwaster> no reason to commit unless it's mentally ill...
<lamont> but yeah once PEG converges the ascent guidance state machine never goes backwards, and if the peg update cycle fails, it restores the old guidance
<lamont> i’m seeing cases where PEG wigs the fuck out though even though it says its converged
<lamont> i think i need to look at how sensitive the convergence check can be made
<lamont> or back the gain off for high angles
<Starwaster> I suspect in my case that a non standard engine might have been responsible... I don't think the SRBs I was using were identifiable as engines. MJ couldn't even provide a proper DV on them
<Starwaster> gotta remember to follow up on that and see what modulet hat engine is using
<lamont> ah yeah, if stage stats are busted then PEG will be busted
<lamont> PEG has to predict the future and needs accurate vehicle info
<Starwaster> but, once I drop the uncompatible engine, shouldn't I be able to reset PEG and have it go from that point forward?
<Starwaster> (after that point DV is accurate)
<Starwaster> hmmm ok weird. the SRB IS standard...
<Starwaster> ok, must be the decoupler. that's gotta be it, it couldn't identify it as a decoupler so it didn't know the SRB could be staged away?
<Starwaster> goddamit its not based on the stock decoupler at all. No wonder MJ can't figure out the stats
<lamont> heh that would do it
<Starwaster> SSTUCustomRadialDecoupler
<Starwaster> SSTU has some cool stuff but there's so much non-standard crap everywhere I look screwing up my favorite mods
<lamont> https://github.com/MuMech/MechJeb2/blob/master/MechJeb2/PartExtensions.cs#L83 probably needs a patch roughly like this one
<lamont> do you know why SSTU wiggles the engines so much on startup?
<Starwaster> no idea, I only have a partial SSTU install. I didn't bother with the engines
<lamont> ah that’s what i’ve got it for
<Starwaster> I would GUESS however that it's a gimbal issue... try scaling gimbal limit down to 25%
<Starwaster> oh.... check and see if SSTU engine gimbals use gimbal rate. I'm not sure MJ knows how to deal with gimbal rate. It probably assumes that gimbals respond instantly
<Starwaster> (which they do by default)
<Starwaster> see, to me, ALL engines wiggle a lot because my game is patched so that all engine gimbals have a gimbal rate
<lamont> this is when they should just be pointed at vessel.forward though
<Starwaster> why?
<Starwaster> clarify?
<lamont> going straight up at launch
<Starwaster> roll?
<lamont> shouldn’t be until 100m or whatever
<lamont> i’ll try to pay more attention to the controls though and see if there’s any commands they’re seeing
<lamont> gimbal rate is an interesting hint
<Starwaster> yes but roll only has to deviate a fraction for MJ to try to bring it back inline with its starting roll position
<Starwaster> you launch, it's a bit off, MJ tries to bring it back
<Starwaster> do that same launch with force roll OFF
<Starwaster> it probably wont do the wiggle
<wb99999999> oh, the Black Arrow
<Starwaster> see, the thing about gimbal rate is that MJ expects to get a certain amount of torque out of gimbals but until the gimbal is fully in position it wont get that expected torque
<lamont> yeah and possibly when the engines aren’t spooled up MJ gives a very large command to what it sees as a very low thrust engine and then it gimbals and doesn’t gimbal back fast enough as it spools up
<lamont> right yeah so it pushes it to the rails of pitch/roll/yaw initially and then the engine is slow to respond back
<Starwaster> maybe it shouldn't look at the current performance but instead base its decisions on rated performance. PID should be enough to correct for that as the errors build up
<lamont> oh that might be the cause of some wigglies i’ve seen as well.
<lamont> yeah although it should drop rated torque down when agathorn strikes
<lamont> having a gimbal sieze would be an interesting failure mode
<Starwaster> that could well be game over, time to abort
<Starwaster> unless it's really close to staging that bit
<Starwaster> wish people wouldn't give me snacks for Xmas.... I swear I can feel my teeth rotting in my head now
<Starwaster> so, I had a feral cat that I was trying to lure onto my front porch and she was really cagey because she's already been trapped twice.... even tuna fish wasn't doing it. Then I put her brother outside in a crate...
<Starwaster> BAM!!!! She runs right up to him to investigate
<Starwaster> then I moved him onto the porch and waited until she came in... pull a string connected to a wooden slat holding the porch door open. SLAM!!!
<lamont> you rehab them?
<Starwaster> it's a sanctuary for feral cats.
<Starwaster> she and her brother have been fed by humans since they were kittens but the whole indoors thing is new to them
<Starwaster> she's not too keen on being petted but she WILL let me pick her up. (she's not real keen on that either)
<Starwaster> her brother loves being petted but doesnt let me pick him up
<Starwaster> originally they were going to be outdoors cats but then things happened and we decided we wanted them in
<Starwaster> he was easy, he just walked in of his own accord
<Starwaster> lamont: ok, MJ DOES look at gimbal rate.... (gimbalResponseSpeed)
Probus__ has quit [Quit: Leaving]
<lamont> interesting. doesn’t necessarily mean its wired up to the PID controllers correctly…
ferram4 has quit [Read error: -0x1: UNKNOWN ERROR CODE (0001)]
blowfish has joined #RO
ferram4 has joined #RO
jclishman has joined #RO
jclishman has left #RO [#RO]
<Starwaster> so it affects TfV... still working my way through t he code from there
ProjectThoth has joined #RO
aradapilot has quit [Ping timeout: 207 seconds]
blowfish has quit [Quit: Leaving]
qwertyy_ has joined #RO
ProjectThoth has quit [Quit: +++out of cheese error+++]
qwertyy has quit [Ping timeout: 186 seconds]
Maxsimal has joined #RO
wb99999999 has quit [Ping timeout: 180 seconds]
Maxsimal has quit [Ping timeout: 180 seconds]
ferram4 has quit [Read error: Connection reset by peer]
ferram4 has joined #RO
TM1978m has quit [Remote host closed the connection]
Majiir has quit [Quit: CUT THE HARDLINES!!]
Snoozee has joined #RO
Snoozee is now known as Majiir
Starwaster has quit [Ping timeout: 383 seconds]
awang has quit [Ping timeout: 383 seconds]
SirKeplan has quit [Ping timeout: 198 seconds]
soundnfury has quit [Ping timeout: 198 seconds]
aradapilot has joined #RO
SirKeplan has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
Starwaster has joined #RO
Senshi has joined #RO
stratochief has quit [Remote host closed the connection]
awang has joined #RO
BadRocketsCo has joined #RO
awang has quit [Quit: leaving]
awang has joined #RO
BadRocketsCo has quit [Ping timeout: 198 seconds]
BadRocketsCo has joined #RO
BadRocketsCo has quit [Ping timeout: 207 seconds]
aradapilot has quit [Ping timeout: 207 seconds]
aradapilot has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
aradapilot has quit [Read error: Connection reset by peer]
aradapilot has joined #RO
aradapilot has quit [Ping timeout: 207 seconds]
aradapilot has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
Bornholio_ has joined #RO
Bornholio has quit [Ping timeout: 200 seconds]
BadRocketsCo has joined #RO
BadRocketsCo has quit [Ping timeout: 198 seconds]
soundnfury has joined #RO
stratochief has joined #RO
awang has quit [Read error: Connection reset by peer]
awang has joined #RO
Senshi has quit [Read error: Connection reset by peer]
wb99999999 has joined #RO
<wb99999999> hey folks
ProjectThoth has joined #RO
wb99999999 has quit [Ping timeout: 180 seconds]
awang has quit [Ping timeout: 198 seconds]
wb99999999 has joined #RO
<wb99999999> Mechjeb/PEG is pretty good :)
<wb99999999> I have a rocket that has 4 RD-275s in a cross arrangment for first stage
<wb99999999> the burn out TWR is high so I'd shut down 2 engines for the last 30s of the stage
<wb99999999> but instead of shutting down one engine each side I shut down 2 engines in the same side
<wb99999999> but PEG finished the burn flawlessly
<wb99999999> good job
awang has joined #RO
<wb99999999> how much more expensive are hypergols?
<wb99999999> I heard that they're significantly more expensive than your kerosene and lox
<Bornholio_> irfna 70$ /liter in smaller quantities http://www.dtic.mil/dtic/tr/fulltext/u2/a036741.pdf crappy USaf doc on NTO Irfna
Bornholio_ is now known as Bornholio
<wb99999999> anything nitric acid is bond to be more costly right?
<Bornholio> depends more on handling requirements, but yes generally
<Bornholio> of course when you purchase rocket fuel amounts the price is much less. One of the larger costs for nasa is dealing with ruined and waste fuels
<Bornholio> that pdf i dropped has costs for the in bulk prices for when it was made
wb99999999 has quit [Ping timeout: 180 seconds]
<Bornholio> say see irfna on page 48 @ $.22/lb for a bulk buy of 600klbs