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"
aradapilot has joined #RO
Senshi has quit [Read error: Connection reset by peer]
ProjectThoth has joined #RO
egg|zzz|egg has quit [Ping timeout: 189 seconds]
TM1978m has joined #RO
stratochief66 has quit [Ping timeout: 378 seconds]
stratochief has quit [Quit: Connection closed for inactivity]
stratochief has joined #RO
Mike` has quit [Ping timeout: 189 seconds]
Mike` has joined #RO
aradapilot has quit [Remote host closed the connection]
stratochief has quit [Quit: Connection closed for inactivity]
ProjectThoth has quit [Quit: +++out of cheese error+++]
Senshi has joined #RO
stratochief66 has joined #RO
egg|zzz|egg has joined #RO
TM1978m has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
stratochief66 has quit [Ping timeout: 378 seconds]
stratochief66 has joined #RO
stratochief66 has quit [Ping timeout: 378 seconds]
Webchat491 has joined #RO
Webchat491 has quit [Quit: webchat.esper.net]
stratochief66 has joined #RO
aradapilot has joined #RO
aradapilot has quit [Remote host closed the connection]
aradapilot has joined #RO
VanDisaster has quit [Quit: Miranda NG! Smaller, Faster, Easier. http://miranda-ng.org/]
VanDisaster has joined #RO
Senshi has quit [Read error: Connection reset by peer]
Senshi has joined #RO
Wetmelon has joined #RO
schnobs has joined #RO
<schnobs> o/
<schnobs> paging Mike`
<Mike`> hey schnobs
<schnobs> Hi, thanks for showing up.
<schnobs> I've just published my "real" data in the issue.
<Mike`> upon quick look i only found those chamber pressure curves in the new pdf btw
<Mike`> ah okay
<schnobs> The "new" and "old" data is available if you want it, but that's nothing I made up -- it's the measured thrust from a test firing in game, using the old and new curves.
<Mike`> any reason you started at -1?
<schnobs> Made it easier for me to get that peak at the beginning.
<schnobs> AFAIK gnuplot doesn't kow about tangents. So in order to get a smooth curve I have to place my points *just right*
<schnobs> That's also the reason why i have that 1400.1 point.
<Mike`> The 1140.1? That's also confusing me (and apparnetly my tool :D), yeah. The reason for that is gnuplot?
<schnobs> basically, yes.
<schnobs> However, you script's output works well in that area. Where it really differs is that it tapers out too fast near the end.
<Mike`> Well, not sutre if that's nwhat you intended, but by going from -1 to 0.2 that's a 1.2 second long "peak"
<Mike`> just replacing the -1 with a 0 is probably more correct?
<Mike`> Then again the peak might well be a second, but the -1 is still confusing me a little :)
<schnobs> Well.... I didn't draw that plot specifically for your script. It was intended for my own reference.
<Mike`> I see
<schnobs> I expected to do a lot of curve tweaking, and overlaying the measured data with that curve would make an easy comparison.
<schnobs> I just fed it into your script to see what happens, and the result is not bad at all :)
<schnobs> Just btw, how do tangents work?
<schnobs> I know what they're supposed to do, but i don't know how to make them.
<Mike`> yeah, with the script you don't have to do too much tweaking usually, you do a somewhat good curve, put it into ksp, and measure the burn time. Either it's spot on or you gotta adjust teither the curve or the thrust a little up/down, or maybe you spot a part of the curve that can be refined a little into the right direction
<Mike`> well, depends what you want to do. in the simple mcase you make tangents linear, which is easy to do, but you only get linear stuff then. If you want somethign else, you need to do some calculation or use a visual editor i assume.
<schnobs> Well, as I understand it it's a statement of "at this point, the curve is supposed to be moving up at such-and-such an angle". Right?
<schnobs> ...or down, or level, but basically I'm giving a direction.
<Mike`> yes
<Mike`> Do you have by chance a verison of you8r data with the brackets and commas added so i can put it into my script?
<schnobs> But how do I declare that angle? Is it simply a vector, 1,1 pointing up 45°?
<Mike`> Want to plot the output and look at it in detail
<Mike`> yes
<Mike`> depending on where you come from, left or right, 1,1 is either 45° up or down
stratochief66 has quit [Ping timeout: 198 seconds]
egg|zzz|egg has quit [Ping timeout: 190 seconds]
<Mike`> so the 1140.1 point you inserted there caused a huge spike, i'm just not sure yet if this is some kind of bug/unwanted behaviour and/or just the way the maths work.
<Mike`> thanks
<Mike`> yep, gnuplot plots the weird spike and it doesn't look good. hm.
<schnobs> What spike?
<Mike`> can't show images here, i'm spoiled by discord now.
<Mike`> but in your curve - this spike:
<Mike`> key = 0.97910 0.95000 -55421.51267 6.23310
<Mike`> key = 0.97900 0.95008 -0.45294 -55421.51267
<schnobs> http://ksp.schnobs.de/RO/plot.png <- my output in full HD
<Mike`> the tangents are huge here, that'S wehy it breaks my formatting but it's also not a good sign
<Mike`> and gnuplot shows that spike doesn't look good indeed when i use it to look at the curve
<Mike`> Yeah you don't see it when zoomed out
<Mike`> you need to zoom to a microscopic level around the spike to see it
<Mike`> :)
<schnobs> Hold on. You're taking about a spike in my input, or the script output?
<Mike`> script output
<schnobs> Does gnuplot know about tangents?
<Mike`> no, but my script outputs a gnuplot formula to plot it
<Mike`> i wonder you didn't notice it as you seem to like gnuplot :)
<schnobs> Maybe because I use gnuplot itself?
<schnobs> A detail around that initial peak, as rendered for me.
ProjectThoth has joined #RO
<Mike`> so what are you plotting now, the outpout of my script or the gnuplot formula? because you won't see it in the output as gnuplot doesn't understand tangents and the spike is in the tangents as we figured out
<Mike`> also it's so small i guess it doesn't matter, but it still bugs me enough to change it. :)
<schnobs> I take the thrust curve from the script, strip off the "key = " parts, then feed it into gnuplot.
<Mike`> yeah you won't see it that way
<Mike`> to make gnuplot see the ncurve like ksp does you need to use the formula my script gives you, you can't use the data.
<schnobs> Hold on, I didn't even notice. it says "Gnuplot: plot x [...]"
<schnobs> I can copy-paste everything from "plot x" onwards into gnuplot?
<schnobs> That gives a short horizontal line with a peak in the middle.
<Mike`> that's how it looks like
<Mike`> yes, you can copy paste that into gnuplot, exactly
<Mike`> then you have to add xrange to restrict it to 0:1
<Mike`> or the range you want to look at
stratochief has joined #RO
<Mike`> that's how the spike looks like, with _really_ large zoom. :)
<schnobs> cannot replicate, my gnuplot chokes at the data (v5.0)
<schnobs> choke at? on?
<Mike`> interesting, what does it do? i just use http://gnuplot.respawned.com/# which works fine
<schnobs> gnuplot> set yrange [0:1]
<schnobs> gnuplot> load "/tmp/rb_plot.txt"
<schnobs> gnuplot> set yrange [0:1]
<schnobs> gnuplot> Segmentation fault
<schnobs>
<schnobs> I *can* plot it, once, and not zoom around too much before it fails.
<schnobs> the web gnuplot refuses to draw anything for me.
<schnobs> but yes, knowing where to look, I can now see the weird spike.
<Mike`> segfault :D okay.
<schnobs> Well anyway, I want to say thanks for that script.
<schnobs> I spent a few hours on sunday drawing up thrust curves. Tweak, load KSP, test, repeat.
<schnobs> This morning, using your script, I gut a much better result instantly, within a whole 15 minutes.
Senshi has quit [Read error: Connection reset by peer]
<Mike`> :D nice to hear. it did take me quite some time to write the script. I started with a primitive version a years ago and greatly improved it a month ago
<Mike`> gotta see when/if i can look into the math to see if that's a bug or something else with the spike
<Mike`> until then i could just do a curve using input better suited for the script.
<schnobs> I've done some limited experiments with tweaking the output of your script, it seems as if most of the tangents aren't even necessary.
<schnobs> I think I can make a curve that will properly fill the hump, though I may have to increase ISP by a few points in order to get there. Let's see if I can get this past the refs; right now, Raidernick seems to like the long tail which I abhor.
Wetmelon has quit [Ping timeout: 183 seconds]
schnobs` has joined #RO
schnobs has quit [Ping timeout: 183 seconds]
<Mike`> schnobs`, "necessary" -- the tangents improve precision, but yeah, depending on how many points you have you might not notice them missing
schnobs` has quit [Ping timeout: 198 seconds]
Shoe18 has quit [Quit: Connection closed for inactivity]