ferram4 changed the topic of #RO to: Welcome to the discussion channel for the Realism Overhaul (meta)mod for KSP! Realism Overhaul Main Thread https://goo.gl/wH7Dzb ! RO Spreadsheet http://goo.gl/Oem3g0 ! Code of Conduct http://goo.gl/wOSv2M ! | Maximal & soundnfury's RP-1 Race Into Space Signup: http://bit.ly/2DEVm2i [15:01] <soundnfury> Straight Eight Stronk (and) RP-0/1 is basically "Space Agency Spreadsheet Simulator"
<Raidernick>
pap is going to convert the ua1205 curve
<Raidernick>
since it's pretty much identical
<Raidernick>
the new one that schnobs just made
<Mike`>
i see
<Raidernick>
i'm not in a rush but pap wants to get it released
<Raidernick>
so it's up to him i just let him know the current one is broken
<Raidernick>
so he doesn't release it with a busted config
<schnobs>
Hold on. My latest PR from Tuesday or so has a 1205 curve that's definitely not broken.
<Raidernick>
schnobs, it's not merged yet
<Raidernick>
if you want to add the new curve to that you can
<Raidernick>
i'm having it be held until the last minute
<schnobs>
I'm still tweaking on the 1207. Might have something presentable in 15 min, hold on.
<Raidernick>
k cool
<schnobs>
However, more PRs will be coming soon(TM) in any event.
<schnobs>
Mike`: Without tangents, I need three data points to describe a bend; with tangents, a single point +t might do. I don't know, I still haven't figured out how to write them.
<schnobs>
So I stick to what I know and can.
<schnobs>
I still take some pride in getting good results with ten data points. The tiny squiggle-wiggles at ignition and tail-off each add just as many data points (and try-rinse-repeat effort), that's why I'm skipping them.
<schnobs>
(ignition peak is easier and probably also more worthwhile to have, I might look into it. But a few sec of +-3% thrust at tailoff really makes no difference)
<schnobs>
Still not quite right, but a great improvement over the current curve.
<schnobs>
Raidernick: ^
<Raidernick>
yeah that's better
<schnobs>
PR incoming.
<schnobs>
done.
<Raidernick>
k i'll check it when i'm done updating my mods
<Raidernick>
uhh
<Raidernick>
what is wrong with that thrrust curve
<Raidernick>
schnobs, it's tiny
<Raidernick>
it's only 11 keys
<Raidernick>
something is wrong with it?
<Raidernick>
also it stays at 100% thrust the entire time
schnobs has quit [Ping timeout: 190 seconds]
egg|zzz|egg has quit [Ping timeout: 189 seconds]
NathanKell|AFK has quit [Ping timeout: 206 seconds]
NathanKell|AFK has joined #RO
Mike` has quit [Ping timeout: 190 seconds]
Mike` has joined #RO
wb99999999 has joined #RO
<wb99999999>
hey guys
wb99999999 has quit [Quit: webchat.esper.net]
stratochief has quit [Quit: Connection closed for inactivity]
Senshi has joined #RO
<travis-ci>
Build #6183 - v12.6.0 - passed
<travis-ci>
Fix the Solar Panel Planner (now only works in VAB)
<travis-ci>
Build #6184 - 1.3.1-FINAL - passed
<travis-ci>
Fix the Solar Panel Planner (now only works in VAB)
schnobs has joined #RO
<Mike`>
hey schnobs. Well, i meant why don't you use the tangents my script provides [as long as you don't trigger the bug you found, need to try to fix that, but i think it doesn't trigger too often]
<Mike`>
With the tangents from my script, you still need quite some points to describe bends, but less points to describe straight lines. Without them you need more points for straight lines as well.
<Mike`>
Also, not providing tangents might not do what you expect. If you don't provide tangents, KSP will add tangents itself, which sometimes do what you want, but sometimes freak out and don't do what you want.
<schnobs>
Hold on. I literally just got up and didn't have a sip of coffee yet...
<Mike`>
:) me neither
TM1978m has joined #RO
<schnobs>
There's a mod that lets you view curves in-game, as KSP sees them. Tried that once and got mad -- I couldn't enter more than three or four data points without accidentally erasing everything.
<schnobs>
That was a few years yago and I haven't looked at it since, maybe it got better idk.
<Mike`>
Yeah i used that as well, it's not the best
<Mike`>
no i don't think so
<Mike`>
That's why i did my script, no more fiddling, it usually just works pretty well. But yeah, need to look at that issue.
<schnobs>
Anyway, I've been using gnuplot ever since. Which doesn't know about tangents, hence I don't use them.
<Mike`>
[Btw, what mod do you use to record ingame thrust values?]
<Mike`>
well, gnuplot not only doesn't know about tangents, it also doesn't know about the propellant domain ksp uses for thrust curves, you cannot compare those to time domain curves you use as input/find on the web.
<schnobs>
krpc.
<Mike`>
k
<schnobs>
While thrust > 0: print thrust; sleep 0.1
<Mike`>
i see
<Mike`>
too bad MJs flight recorder doesn't have it
<schnobs>
krpc is pretty lightweight, easy to use, and of course gives you insane freedom.
<schnobs>
IMO the main strength lies in the small things. You can of course put together intricate ascent scripts... but what's really valuable as that if you find yourself having a particular problem, you can easily whip up a ten-line quick fix.
<Mike`>
yeah, might try that some time
<schnobs>
Like starting the burn not 1/2 burn time ahead of the node, but 1/2 impulse. Makes quite a difference on centaur.
<Mike`>
i'm using principia, no more stock nodes madness. :)
<Mike`>
[Which i can heavily recommend btw, it's eggscelent :]
<schnobs>
Anyway, this was my first curve, and until now the onyl time I used your script.
<schnobs>
I'd have been happy to just hand in the curve as it came out, but when it was determined that it needs a workover, I started to tweak it by hand and the tangents were the first to go.
<schnobs>
IIRC I could just toss out the tangents and would still get the very same result in-game from just having the data points.
<schnobs>
Not 100% sure.
<schnobs>
And thrust-propellant curves can be tweaked by hand, no problem. The way they're distored realtive to thrust-time is a bit irritating, of course.
<schnobs>
But I already had a pretty decent thing to start with, only *here* it needed to last a little longer, and had to drop of faster *there*... that's easy.
<schnobs>
In the olden days, I'd arrive at that point by counting boxes on the grid. I'd need a dozen iterations, at least, to get as close as with the first pass of your script.
<Mike`>
the tangents just improve accuracy, depending on how close the points are, you might not notce them at all, yeah, but the fewer points there are, the greater the difference by the tangents is
<Mike`>
but i see. I usually just try to fine tune the input to my script a little if there's still some area that needs tuning
<schnobs>
do tangents have a scale? Does it make a difference whether I write 1,5 or 100, 500?
<schnobs>
And yes, it's likely a personal thing. Fine-tuning the curve give me the feeling of having done my part :)
<Mike`>
the numbers describe 2 different tangents. And really, i don't expect anyone to do tangents by hand unless you're doing straight line tangents or use a GUI to do them, so yeah, i wouldn't bother doing those manually for the curves.
<Mike`>
I just thought you used my script but stripped them off which didn't make a whole lot of sense to me that's why i asked
<Mike`>
But at least if i manage to fix that small issue i'd suggest to try fine tuning the input of my script as well and see how that works, because i'd hope that to be easier/less work as you can work without having to worry about the distortion. Also if you have suggestions for improvement, i'd also like to hear them :)
<schnobs>
not sure if improvement to your script, as such, but I wondered if there's a way to get interpolated data out of gnuplot.
egg|zzz|egg has joined #RO
<schnobs>
I give it three points, it draws a nice bend... now could it hand me a list of all the x,y values it just plotted?
<schnobs>
That list could be fed into a simulator, so I could again arrive at the resulting thrust:time curve without having to start KSP.
<schnobs>
I guess a possible solution, or pointers to a solution, are somewhere in your script. You're also creating curves from points, after all.
<Mike`>
Well, my script only does straight lines in the time domain, just uses the input you give it, but on the other hand i think that's fine for srm thrust curves - some really have hard corners, but even for smooth curves irl, nobody's really gonna notice not entirely smooth corners i guess.
schnobs has quit [Ping timeout: 206 seconds]
Daz has quit [Ping timeout: 183 seconds]
Kerbnub has joined #RO
Kerbnub has quit [Client Quit]
schnobs has joined #RO
egg|zzz|egg has quit [Ping timeout: 190 seconds]
egg|zzz|egg has joined #RO
ferram4_ has joined #RO
ferram4 has quit [Ping timeout: 198 seconds]
egg|zzz|egg has quit [Ping timeout: 198 seconds]
egg|zzz|egg has joined #RO
Senshi has quit [Read error: Connection reset by peer]