[0.25]KSP Interstellar (Magnetic Nozzles, ISRU Revamp) Version 0.13

[Lightwarrior]

Decided to test performance. Copied KSP to my old PC (Phenom2 x4 3.2Ghz, 8gb DDR2, GF9600GT), loaded large enough craft (~350part), and... it works almost the same as stock game. ~10FPS.

But there is one interesting thing i noticed - those problems with EC/MJ and fusion reactors shuting down at high warp are... gone. No problems at all. EC/MJ stays full even at 100.000x and fusion reactors continue to work just fine. Seems strange that bad performance fixed something...

[Fractal_UK]

Decided to test performance. Copied KSP to my old PC (Phenom2 x4 3.2Ghz, 8gb DDR2, GF9600GT), loaded large enough craft (~350part), and... it works almost the same as stock game. ~10FPS.

But there is one interesting thing i noticed - those problems with EC/MJ and fusion reactors shuting down at high warp are... gone. No problems at all. EC/MJ stays full even at 100.000x and fusion reactors continue to work just fine. Seems strange that bad performance fixed something...

Did this occur using an identical craft to one previously tested on another machine or did the design differ?

[Lightwarrior]

I just made full copy of KSP folder, same save, same craft etc.

//add

Experimented with it a bit more and noticed that timewarp rate changes really slow here. ~2 game days passes before it can be changed from 100.000x to 1x. May be this solves those resource management problems...

Edited December 12, 2013 by Lightwarrior

[NotCoach]

I was trying to optimize my DT Vista setup last night, and I ended up doing a config cheat on the largest generators to run at maximum efficiency, and I'm wondering if it would be possible to implement what I was aiming for into the mod.

I wanted to maximize each rector's potential in order to have the most possible DT Vistas per reactor. Using a single reactor (thorium fueled) and generator (both 3.75), and assuming 58.7% efficiency on the generator, one can power ~2.47 DTs. But what about 2 generators per reactor? No difference, of course, due to how the mod works. Thermal power is just simply shared between the two reactors, thus halving the output of each. Extra weight with no extra benefit. But instead of just simply sharing the thermal power why not prioritize it? What I mean is why not send all the thermal power to generator 1, then shunt the wasted heat to generator 2? Continuing to assume 58.7% efficiency the first generator produces 6.046 GWe with 4.254 GWt wasted heat. Normally that would be sent to the radiators and lost forever. What if, instead, that left over thermal power is shunted to the second generator? The second generator would produce 2.497 GWe, and 1.757 GWt would be lost instead. And instead of using 58.7% of the thermal power we would be using ~82.9% of the reactors thermal power.

The output rate of each generator is reduced, of course. This is the tradeoff. Generators are lighter than reactors, so it is preferable to run two generators per reactor with the above outputs. I would rather produce 4.272 GWe per generator using only 1 reactor instead of producing 6.046 GWe per generator with only 1 generator per reactor. So is it possible to incorporate this into the mod? As I said above I am emulating it by cheating the generator config (raising efficiency until I get 4.272 GWe per generator on a one reactor, 2 generator setup).

[forsaken1111]

Yeah I agree that the detailed 'accuracy' does not matter. For most applications, the base passive radiation will be insignificant compared to the total radiation needed. Personally, I think that the base passive radiation should be high enough for a small satellite with a couple of standard extending panels to radiate sufficiently to not need radiators. Once you move on to bigger things, the passive radiation will be insignificant to your needs anyways. On the other hand, perhaps you like it the way it is, where radiators are needed pretty much from the beginning.

This mod is based on realistic or theoretically possible technology unfortunately, and even small satellites need some help cooling down in the harsh light of the sun. Most craft in space simply do not radiate heat efficiently unless they are designed to do so.

The good news is that a single tiny radiator can handle the heat from a few panels.

[Fractal_UK]

Speaking of radiators - good news for B9 Aerospace players in 0.9. I've decided to finally implement B9 cargo bay doors acting as radiators - based on the current cargo bay doors, because they are using a non-stock animation module I cannot make them make them like deployable radiators that work only when the doors are open, this is a shame but it may be that this will be possible in the future depending on what 0.23 brings and the changes that bac9 makes as a result of it.

I'd never really considered building a 3.75m antimatter powered aircraft before but actually it works rather well, the largest reactor has poor TWR compared to the smaller ones but good enough specific impulse that using LiquidFuel all the way to orbit isn't really a problem and the lift from the wings helps to offset that lower TWR. I've built a few rockets using those 3.75m reactors but this might be a better way of using them on the ground.

Of course, it's among the latest of late game things you could do but still a potentially nice cargo platform.

This is all very appropriate because version 0.9 is set to contain quite a few changes with regard to the aerospace side of KSP, so those of you who are interested in spaceplanes and the like should enjoy the update.

So far, I've also managed to coax a fair bit more atmospheric performance out the thermal turbojets, the changes are:

1. Velocity curve has been tweaked to affect performance almost exclusively at high speed rather than affecting the low speed regime - this means that all thermal turbojets will have more thrust at takeoff and make those fission powered jets a bit more flyable. This is a needed change because unlike stock engines, my engines also perform worse in the low atmosphere due to lower specific impulse so the current system is like a double penalty.
   
2. The atmospheric limit code no longer tops out at ~80-85%, which again means more thrust for everything that flies with a thermal turbojet.
   
3. Detection of flameout has been considerably improved so that the engines aren't hampering themselves when more thrust should be available.
   

[Donziboy2]

Speaking of radiators - good news for B9 Aerospace players in 0.9. I've decided to finally implement B9 cargo bay doors acting as radiators - based on the current cargo bay doors, because they are using a non-stock animation module I cannot make them make them like deployable radiators that work only when the doors are open, this is a shame but it may be that this will be possible in the future depending on what 0.23 brings and the changes that bac9 makes as a result of it.

I'd never really considered building a 3.75m antimatter powered aircraft before but actually it works rather well, the largest reactor has poor TWR compared to the smaller ones but good enough specific impulse that using LiquidFuel all the way to orbit isn't really a problem and the lift from the wings helps to offset that lower TWR. I've built a few rockets using those 3.75m reactors but this might be a better way of using them on the ground.

Of course, it's among the latest of late game things you could do but still a potentially nice cargo platform.

This is all very appropriate because version 0.9 is set to contain quite a few changes with regard to the aerospace side of KSP, so those of you who are interested in spaceplanes and the like should enjoy the update.

So far, I've also managed to coax a fair bit more atmospheric performance out the thermal turbojets, the changes are:

1. Velocity curve has been tweaked to affect performance almost exclusively at high speed rather than affecting the low speed regime - this means that all thermal turbojets will have more thrust at takeoff and make those fission powered jets a bit more flyable. This is a needed change because unlike stock engines, my engines also perform worse in the low atmosphere due to lower specific impulse so the current system is like a double penalty.
   
2. The atmospheric limit code no longer tops out at ~80-85%, which again means more thrust for everything that flies with a thermal turbojet.
   
3. Detection of flameout has been considerably improved so that the engines aren't hampering themselves when more thrust should be available.
   

Hmm twin orange tanks to space SSTO here I come And i'm talking without AM reactor:P

[wasmic]

Hey Fractal-UK, I've found a rather serious bug in this mod. This mod makes any pod with a build-in decoupler not work. The symptoms are varying: It can be everything from the pod auto-decoupling at physics initialization to everything other than the pod disappearing upon load. The pods that I have exprienced this with are the Odin pod from NovaPunch and the Dragon pod by CBBP. The only thing in common between these are that they have a build-in decoupler, and that they only stopped working when I installed this mod.

[Myrten]

Is this WAD that solar panels few hundred kilometers from Sun's surface only produce 10x more energy then on launchpad?

[Doomydoom]

Is this WAD that solar panels few hundred kilometers from Sun's surface only produce 10x more energy then on launchpad?

I hope not, it really crushed my dreams of a solar farm smaller than hundreds of parts.

[Donziboy2]

Is this WAD that solar panels few hundred kilometers from Sun's surface only produce 10x more energy then on launchpad?

I thought that was handled by stock KSP not KSPI. I have a stock install on my labtop, I will hit it up after work.

[Myrten]

Interesting thing is that their output has stopped increasing at some distance around a sun (like 1 million km) just like there would be some kind of a cap.

[Donziboy2]

Interesting thing is that their output has stopped increasing at some distance around a sun (like 1 million km) just like there would be some kind of a cap.

Its stock, ran it on laptop with stock game with only hyperedit to get to sun, it was 10x output at 460km.

[Myrten]

Its stock, ran it on laptop with stock game with only hyperedit to get to sun, it was 10x output at 460km.

I think something is wrong in code, let me present my case:

                    double output = panel.flowRate;
                    double spower = part.RequestResource("ElectricCharge", output * TimeWarp.fixedDeltaTime);
                    double inv_square_mult = Math.Pow(Vector3d.Distance(FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBIN].transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2) /          Math.Pow(Vector3d.Distance(vessel.transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2);
                    displayed_solar_power += spower / TimeWarp.fixedDeltaTime;
                    solar_power += spower / TimeWarp.fixedDeltaTime/inv_square_mult;

displayed_solar_power does not include inv_square_mult which takes distance to sun into account so while it's added to actual transmitted value it's not displayed hence it looks like stock.

So I've changed that line into this to see what will happen:

                    displayed_solar_power += spower / TimeWarp.fixedDeltaTime / inv_square_mult;

Now displayed beamed power is equal to what is actually transmitted and result is quite surprising...

The closer you are to sun the less power is produced...

So I've changed last two lines into:

displayed_solar_power += spower / (TimeWarp.fixedDeltaTime / inv_square_mult);
solar_power += spower / (TimeWarp.fixedDeltaTime/inv_square_mult);

I'm not sure if that was Fractal's goal but with this change it seems to work well

Edited December 12, 2013 by Myrten

[Donziboy2]

Myrten, your missing the point, it is stock KSP functionality. Fractal_UK did not change the amount panels produce at different distances.

And also expecting 10 times the output from solar panels in RL would probably destroy the panels. Let alone 1000000 times the output.

[Myrten]

Myrten, your missing the point, it is stock KSP functionality. Fractal_UK did not change the amount panels produce at different distances.

And also expecting 10 times the output from solar panels in RL would probably destroy the panels. Let alone 1000000 times the output.

His original code does indeed changes transmitter output, it's just not displayed in 'Beamed power'.

double inv_square_mult = Math.Pow(Vector3d.Distance(FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBIN].transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2) /          Math.Pow(Vector3d.Distance(vessel.transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2);

This variable store Distance(Kerbin,Sun)² divided by Distance(Vessel,Sun)²

It's then used in solar_power += spower / TimeWarp.fixedDeltaTime/inv_square_mult; which is not displayed on transmitter but should be visible at receiver.

You can check that by placing two vessels, one with solar panels + transmitter and second one with receiver pointed at first vessel.

At Kerbin it will transmit ~100 KW and receiver will get that, now if you move these two vessels to low Sun orbit receiver will get less power despite the fact that transmitter still displays ~100 KW as display at transmitter is IMHO bugged as well as order of operations which results in less power closer to sun.

[Fractal_UK]

I think something is wrong in code, let me present my case:

                    double output = panel.flowRate;
                    double spower = part.RequestResource("ElectricCharge", output * TimeWarp.fixedDeltaTime);
                    double inv_square_mult = Math.Pow(Vector3d.Distance(FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBIN].transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2) /          Math.Pow(Vector3d.Distance(vessel.transform.position, FlightGlobals.Bodies[PluginHelper.REF_BODY_KERBOL].transform.position), 2);
                    displayed_solar_power += spower / TimeWarp.fixedDeltaTime;
                    solar_power += spower / TimeWarp.fixedDeltaTime/inv_square_mult;

displayed_solar_power does not include inv_square_mult which takes distance to sun into account so while it's added to actual transmitted value it's not displayed hence it looks like stock.

You are mistaken, there is nothing wrong with this code. This code gets a raw power value from the solar panels and displays it - there is no inverse square multiplier quite deliberately. The solar panels already take distance into account themselves so why would there be another inverse square factor displayed?

The solar power part, however, contains the inv_square_mult because it is designed to scale the energy that will be stored in the microwave persistence code to what it would be at Kerbin orbit. That is why dividing by the inv_square_mult is giving you a smaller value. This is because there is no guarantee that the satellite will be at the same place, in the future, as it was when you turned the transmitter on so Kerbin is used as a common reference point and the receiver will scale the energy based on the transmitter's current position.

The problem you have detected is that the panels are indeed using a stock curve, rather than the corrected curve that I included - I am not entirely sure why this is the case but that is why you're seeing 10x the power in the sun's orbit because 10x is the maximum amount of power that the stock panels can ever produce. The problem is with the solar panels not the transmitter, the transmitter code is fine.

[Myrten]

You are mistaken, there is nothing wrong with this code. This code gets a raw power value from the solar panels and displays it - there is no inverse square multiplier quite deliberately. The solar panels already take distance into account themselves so why would there be another inverse square factor displayed?

The solar power part, however, contains the inv_square_mult because it is designed to scale the energy that will be stored in the microwave persistence code to what it would be at Kerbin orbit. That is why dividing by the inv_square_mult is giving you a smaller value. This is because there is no guarantee that the satellite will be at the same place, in the future, as it was when you turned the transmitter on so Kerbin is used as a common reference point and the receiver will scale the energy based on the transmitter's current position.

The problem you have detected is that the panels are indeed using a stock curve, rather than the corrected curve that I included - I am not entirely sure why this is the case but that is why you're seeing 10x the power in the sun's orbit because 10x is the maximum amount of power that the stock panels can ever produce. The problem is with the solar panels not the transmitter, the transmitter code is fine.

OK, now I got how it's supposed to work . Is your corrected curve also supposed to be based on Distance(Vessel,Sun)² divided by Distance(Kerbin,Sun)² ? Everything would be logical then as first it would scale flow to the current distance, then store value unscaled by distance in Microwave persistence and finally scale it again in receiver based on transmitter's current location.

[Fractal_UK]

OK, now I got how it's supposed to work . Is your corrected curve also supposed to be based on Distance(Vessel,Sun)² divided by Distance(Kerbin,Sun)² ? Everything would be logical then as first it would scale flow to the current distance, then store value unscaled by distance in Microwave persistence and finally scale it again in receiver based on transmitter's current location.

Yes, to be honest, I don't know why that isn't working. Are your solar panels generating waste heat properly? The change is made from that PartModule, so if the solar power curve isn't working, it's quite likely solar waste heat won't be either.

My Low Kerbol Orbit test satellite is at a similar altitude to your solar satellite. The stock solar panels start at 18KW generation, here I'm getting 7MW per panel, which is 389x the expected Kerbin output. Taking into account my altitude of 417,834,000m (and Kerbol's 261,600,000m radius) and Kerbin's orbital radius of 13,599,840,256, we expect (13,599,840,256/(417,834,000+261,600,000))^2 = (20.01..)^2 = 400.65x the output at Kerbin orbit - which is very similar to my 389x figure I'm achieving in practice, the difference can easily be explained by my solar panels being not perfectly aligned.

Edited December 13, 2013 by Fractal_UK

[Donziboy2]

Yes, to be honest, I don't know why that isn't working. Are your solar panels generating waste heat properly? The change is made from that PartModule, so if the solar power curve isn't working, it's quite likely solar waste heat won't be either.

My Low Kerbol Orbit test satellite is at a similar altitude to your solar satellite. The stock solar panels start at 18KW generation, here I'm getting 7MW per panel, which is 389x the expected Kerbin output. Taking into account my altitude of 417,834,000m (and Kerbol's 261,600,000m radius) and Kerbin's orbital radius of 13,599,840,256, we expect (13,599,840,256/(417,834,000+261,600,000))^2 = (20.01..)^2 = 400.65x the output at Kerbin orbit - which is very similar to my 389x figure I'm achieving in practice, the difference can easily be explained by my solar panels being not perfectly aligned.

Then I would have to say its been broken for a while, I get the same output for both 0.8.2.1 and the Dev Build.

3.969MW x 16 = 63.504MW

I wasn't aware you had used any multipliers so I assumed it was normal.

[Myrten]

Yes, to be honest, I don't know why that isn't working. Are your solar panels generating waste heat properly? The change is made from that PartModule, so if the solar power curve isn't working, it's quite likely solar waste heat won't be either.

My Low Kerbol Orbit test satellite is at a similar altitude to your solar satellite. The stock solar panels start at 18KW generation, here I'm getting 7MW per panel, which is 389x the expected Kerbin output. Taking into account my altitude of 417,834,000m (and Kerbol's 261,600,000m radius) and Kerbin's orbital radius of 13,599,840,256, we expect (13,599,840,256/(417,834,000+261,600,000))^2 = (20.01..)^2 = 400.65x the output at Kerbin orbit - which is very similar to my 389x figure I'm achieving in practice, the difference can easily be explained by my solar panels being not perfectly aligned.

Something is really weird here...

I've removed all my today's changes (made after discovering initial bug so it's not their fault) to transmitter code and recompiled it, launched game again and it seems to be working now (maybe vessel has to be leaved and re-activated for new curve to work???), but there is another bug...

Beamed power is equal to flow of single panel only instead of their sum.... But on the other hand generated waste heat is correct...

Edited December 13, 2013 by Myrten

[SeventhArchitect]

Sup guys! Just popping in right now to check on the progress of this mod and it seems that it's gotten a lot more stuff since I lost my pc. Might not be important for any of you to know, but, I mean, I do have bragging rights. Anyone who knows me will say the same, so when I say the architect's out of commission that means I'm stuck without building something epic.

Anyway, ego and stuck up crap out of the way, the only thing I would question here is if the AM and nuclear parts got that new model they desperately needed, and if the whole wobble BS that happens a lot got fixed. I know its kind of an asshat thing to say but everyone would definitely agree that when that damned singular AM tank, attached to 24 struts, being one of the more important pieces of the ship, suddenly derps and shrugs while it slowly falls away from the launching rocket, it gets a little old and annoying. Plus, I mean, simply the ffactthat these are big pieces should contribute to the bit of the laws of physics where it should not wobble.

Anything on a wormhole drive yet?

Anything on massive power generation after AM power gen?

Anything in particular I should know about because I missed it and it is crucial to the game and/or just really big in terms of an addition?

[Eadrom]

Anyway, ego and stuck up crap out of the way, the only thing I would question here is if the AM and nuclear parts got that new model they desperately needed, and if the whole wobble BS that happens a lot got fixed. I know its kind of an asshat thing to say but everyone would definitely agree that when that damned singular AM tank, attached to 24 struts, being one of the more important pieces of the ship, suddenly derps and shrugs while it slowly falls away from the launching rocket, it gets a little old and annoying. Plus, I mean, simply the ffactthat these are big pieces should contribute to the bit of the laws of physics where it should not wobble.

Ferram came out with a plugin that provides high levels of stability on your rocket stacks. Basically, the larger the parts, the tighter they stick to each other. I rarely/never have wobble issues since starting to use it. Here is a link: http://forum.kerbalspaceprogram.com/threads/55657-0-22-Kerbal-Joint-Reinforcement-v1-4-2-Properly-Rigid-Part-Connections

Anything on massive power generation after AM power gen?

Upgraded AM reactors generate absolute gobs of power, but if you want MOAR, you can set up a constellation of microwave beaming satellites, either solar arrays in low Kerbol orbit or large fission/fusion power stats. If you invest enough in setting up the infrastructure, you get get ridiculous amounts of power on your vessels.

Anything in particular I should know about because I missed it and it is crucial to the game and/or just really big in terms of an addition?

Are you referring to KSP Interstellar or just KSP in general? For KSPI, check out the FAQ and read the patch notes (see the OP). For general KSP news, read the patch notes and be sure to check out the KerbalKon streams from today and tomorrow.

Edited December 13, 2013 by Eadrom

[simmy2109]

Hey Fractal-UK, I've found a rather serious bug in this mod. This mod makes any pod with a build-in decoupler not work. The symptoms are varying: It can be everything from the pod auto-decoupling at physics initialization to everything other than the pod disappearing upon load. The pods that I have exprienced this with are the Odin pod from NovaPunch and the Dragon pod by CBBP. The only thing in common between these are that they have a build-in decoupler, and that they only stopped working when I installed this mod.

Interesting that you point this out... I was having similar strange issues with the NP Odin pod. I haven't bothered to test it much to determine a cause. Is there any reason to think that KSPI could cause these strange behaviors?

[Fractal_UK]

Something is really weird here...

I've removed all my today's changes (made after discovering initial bug so it's not their fault) to transmitter code and recompiled it, launched game again and it seems to be working now (maybe vessel has to be leaved and re-activated for new curve to work???),

The curve is updated every frame at the same time as the WasteHeat display.

but there is another bug...

Beamed power is equal to flow of single panel only instead of their sum.... But on the other hand generated waste heat is correct...

You can see in my picture on the previous code that 4 solar panels are working correctly in terms of energy production. Can you try an install of just the base plugin dll and see if it works properly then?

Anyway, ego and stuck up crap out of the way, the only thing I would question here is if the AM and nuclear parts got that new model they desperately needed

Antimatter parts have had new models for ages, the nuclear ones arrived in 0.8 - you can see a picture of all 4 sizes of nuclear reactor in the first thread.

and if the whole wobble BS that happens a lot got fixed. I know its kind of an asshat thing to say but everyone would definitely agree that when that damned singular AM tank, attached to 24 struts, being one of the more important pieces of the ship, suddenly derps and shrugs while it slowly falls away from the launching rocket, it gets a little old and annoying. Plus, I mean, simply the ffactthat these are big pieces should contribute to the bit of the laws of physics where it should not wobble.

It's fixed to about the limit it's possible to fix - it's possible to fly a 3.75m rocket stack into orbit without any struts at all but I can't guarantee that parts will be totally wobble free, KSP and big parts don't get along all that well. I have heard good things about ferram's joint reinforcement plugin, a lot of people over at the Real Sized Solar System are having good results with realistically sized rockets, which can be much larger than 3.75m.

Interesting that you point this out... I was having similar strange issues with the NP Odin pod. I haven't bothered to test it much to determine a cause. Is there any reason to think that KSPI could cause these strange behaviors?

Unfortunately it seems he's right. It's caused by the module that determines capsule radiation, I'll fix this problem for the next update.