[WIP] Nert's Dev Thread - Current: various updates

[Captain Sierra]

20 minutes ago, Starbuckminsterfullerton said:

If you want to get really fancy Nert, you could make the boiloff rate a function of the tank temperature, and require radiators/insulators or power to keep them cool and avoid transferring heat generated in other parts of the ship.  For short trips like launches you wouldn't need to bother, but for longer voyages the extra mass would start to pay off.

I already proposed this in a suggested implementation which I've quoted below for your reading pleasure.

On 12/25/2015 at 10:17 PM, Captain Sierra said:

Possible implementation B (complex/thermal version):

Boiloff is per-tank, as above.
Boiloff is a dynamic rate as a function of tank interior temperature (lifter tanks have poor (high) skin-int conduction multipliers, while orbital tanks have lower mults).
Lifter tanks lack any form of refrigeration, boiloff is unpreventable.
Orbital tanks require minimal Ec to refrigerate themselves, reducing interior temperature but increasing core temperature (radiators can pull the core heat away).
As above, lifter tanks have worse mass ratio but better volume ratio (heavier structure but no space lost to insulation/refrigeration equipment), and orbital tanks have better mass ratio but worse volume (space lost to insulation & refrigeration systems, but lighter materials not intended to support 120 ton rockets).

To which was replied the following:

On 12/25/2015 at 11:06 AM, Nertea said:

A is what we will test (I hate the heat system with a passion right now), but with one change:

Same volume ratio, lifter tanks have better mass ratio though.

Anything involving thermal system, while arguably a positive gameplay change, has been ruled out due to Nertea being sufficiently annoyed with it post-NFE overhaul (no. 3). I'm gonna respect his judgement on that.

[Starbuckminsterfullerton]

Thanks Sierra, I try to read a couple pages back but I've been away for a week and have missed stuff, so that helps me catch up. 

And yeah, I remember how long it took to balance NFE with the heat system, way too much work to go through all of that again!

I think I agree with constant volume ratio, since it is adjusted from the real value having different values for different tanks would be a big balancing headache. 

[Nertea]

Initial very rough test, mostly provided for tests of the way the framework interacts. All balance numbers are very not final, and I do appreciate input. I put about 5 minutes of thought into everything.

Contains technically 3 testable subsets, CryoEngines (mostly unchanged from release version), CryoTanks (contains fuel switch patches, foil tanks) and KerbalAtomics (contains NTR patches). 

https://www.dropbox.com/s/85m8zr9225uqquh/CryoStuff_X1.zip?dl=0

[Kerbas_ad_astra]

Some immediate feedback and a heads-up for testers, just from inspecting the zip: there's a bunch of extra DLLs in the CryoTanks plugin folder.  I assume that SimpleBoiloff.dll is the only one of interest.  (Also, you should include or link to the source, per the addon posting rules.)

[Nertea]

Forgot to clean up all the dependency exports I guess. It's harmless but you can delete them if you really want to.

Source is in the same place as always: https://github.com/ChrisAdderley/

[Kerbas_ad_astra]

I thought I had looked there and not found it, but there it is.  Thanks for the heads-up!

[Psycho_zs]

Chelyabinsk does not extend its nozzle when activated.

[Psycho_zs]

...so, boiloff is not linked to temperature or tank volume? I wonder if it is possible to set initial part temperature at launch.

It seems pretty slow. After a week on the launchpad the rocket is still able to go into space, just about 1% of the fuel is gone.

 

[Captain Sierra]

1 hour ago, Psycho_zs said:

...so, boiloff is not linked to temperature or tank volume? I wonder if it is possible to set initial part temperature at launch.

It seems pretty slow. After a week on the launchpad the rocket is still able to go into space, just about 1% of the fuel is gone.

Given the previous discussion, its doubtful you'll notice significant losses after a week. Its meant to make anything interplanetary sufficiently awkward while not hindering any kind of Kerbin SOI applications.

I need to run my own tests (hindered by work, migrane, and Monster Jam) but I suspect the current rate is in a good place.

[billkerbinsky]

2 hours ago, Psycho_zs said:

It seems pretty slow. After a week on the launchpad the rocket is still able to go into space, just about 1% of the fuel is gone.

If we wanted to model it like how things work in real life, you wouldn't worry about boiloff on the pad -- launch clamps would keep tanks topped off until just before launch.

[Fraz86]

@Nertea A few quick notes:

• The basic mechanics of boiloff appear to work as expected. Fuel loss proceeds exactly as predicted for exponential decay with a rate of 0.025% per hour.
• When transitioning into or out of 10000x or 100000x time warp, ZBO tanks often lose a little fuel despite more than adequate solar panels and batteries.
• All ZBO tanks have the same EC requirement (0.25 EC/s). It seems like this requirement ought to scale with the size of the tank.
• As noted in my previous post, 0.025% loss per hour is probably too low to make round trips to Duna using LH2/OX truly awkward. A 7% increase in required fuel mass isn't terribly significant, and thus LH2/OX engines with atmospheric tanks will likely be superior to LF/OX for some Duna missions. Depending on the mass ratio advantage of atmospheric LH2/OX tanks, they might also be superior to ZBO tanks for some Duna missions. I'm afraid these considerations might be confusing/non-transparent to players. I believe the boiloff rate should be at least doubled, in order to make atmospheric more clearly undesirable for interplanetary missions.
• It looks like there haven't yet been any changes to LH2 fudge factor or dry mass for atmospheric or ZBO tanks. Do you have some targets in mind?
• The Poseidon Atomic Rocket's description mentions a built-in generator; looks like that hasn't been implemented yet?

Edited January 9, 2016 by Fraz86

[riocrokite]

Hey guys, does this boiloff plugin work in the background for not loaded vessels? I'm asking because AFAIR the other boiloff plugin (from RO) affected only one selected vessel

[Captain Sierra]

7 hours ago, Fraz86 said:

@Nertea A few quick notes:

• The basic mechanics of boiloff appear to work as expected. Fuel loss proceeds exactly as predicted for exponential decay with a rate of 0.025% per hour.
• When transitioning into or out of 10000x or 100000x time warp, ZBO tanks often lose a little fuel despite more than adequate solar panels and batteries.
• All ZBO tanks have the same EC requirement (0.25 EC/s). It seems like this requirement ought to scale with the size of the tank.
• As noted in my previous post, 0.025% loss per hour is probably too low to make round trips to Duna using LH2/OX truly awkward. A 7% increase in required fuel mass isn't terribly significant, and thus LH2/OX engines with atmospheric tanks will likely be superior to LF/OX for some Duna missions. Depending on the mass ratio advantage of atmospheric LH2/OX tanks, they might also be superior to ZBO tanks for some Duna missions. I'm afraid these considerations might be confusing/non-transparent to players. I believe the boiloff rate should be at least doubled, in order to make atmospheric more clearly undesirable for interplanetary missions.
• It looks like there haven't yet been any changes to LH2 fudge factor or dry mass for atmospheric or ZBO tanks. Do you have some targets in mind?
• The Poseidon Atomic Rocket's description mentions a built-in generator; looks like that hasn't been implemented yet?

Re EC: Oh dear. I know this crowd of people. We love complexity and if you give us an inch, we'll take a mile. Give us EC usage and before you know it, people will be asking for full thermal sim. @Nertea you're the one who gets to draw the line on how much feature bloat is acceptable feature bloat.

Boiloff rate: Based on the data you displayed in that thread Fraz, I concluded that an arguably ideal rate to make Duna one way sufficiently awkward and also difficult was 0.065%. That, however was based on when the model was linear rather than exponential decay. With the exponential model it could go as high as 0.1% to be reasonable. I would definitely advise targeting 0.08% currently.

[Fraz86]

7 hours ago, Captain Sierra said:

Re EC: Oh dear. I know this crowd of people. We love complexity and if you give us an inch, we'll take a mile. Give us EC usage and before you know it, people will be asking for full thermal sim. @Nertea you're the one who gets to draw the line on how much feature bloat is acceptable feature bloat.

I am absolutely not asking for a thermal sim. I think the complexity level is pretty much perfect right where it is. The current implementation is consistent with the middle complexity option from my original boiloff suggestions, which was my personal favorite of the three, so I'm quite happy with it.

EC usage scaling with tank size is really just a matter of consistency and intuitive game mechanics. Consider the following:

• In vanilla KSP, one large tank has identical properties to two half-size tanks, except for modest cost savings. Identical EC consumption for all LH2 tanks (regardless of size) adds a new consideration: larger tanks are much more EC efficient. Thus, non-scaling EC consumption actually increases complexity, because it incentivizes the consolidation of a craft's LH2 to the fewest number of tanks possible - a design consideration not present in vanilla KSP - in order to minimize EC consumption. If EC consumption is proportional to tank size, players can select whatever set of tanks they want (e.g., multiple smaller tanks instead of one big tank) without incurring any disadvantage beyond slightly higher cost, just like vanilla KSP.
• EC consumption for LH2 tanks represents active refrigeration, just like the EC consumption of radiators. The EC consumption of radiators scales with their size.
• It's common sense that refrigerating 10000L of LH2 should require more electricity than 100L, and players will intuitively expect this to be the case.

Quote

Boiloff rate: Based on the data you displayed in that thread Fraz, I concluded that an arguably ideal rate to make Duna one way sufficiently awkward and also difficult was 0.065%. That, however was based on when the model was linear rather than exponential decay. With the exponential model it could go as high as 0.1% to be reasonable. I would definitely advise targeting 0.08% currently.

The data I displayed in that chart was for exponential decay, not linear. It's difficult to make an exact recommendation regarding the final boiloff rate until we know how large the dry mass ratio advantage of atmospheric tanks will be, but - as stated in my previous post - I agree that the rate should be at least doubled (0.05%). Perhaps as much as quadrupled (0.1%), as you said.

Edited January 9, 2016 by Fraz86

[Fraz86]

@Nertea Another note:

In some situations, it may be important to have a manual on/off toggle for refrigeration on ZBO tanks. For example, a Duna ascent vehicle with cryo engines and ZBO tanks probably won't have enough batteries to keep the tanks powered through Duna's 9 hour nights. It would be desirable to turn off refrigeration at night so the craft doesn't run out of power, and the fuel loss to boiloff would be trivial.

[Nertea]

As I mentioned, this is mostly a functionality test. The numbers I used are just testing afterthoughts. I'm more interested in stuff that broke, because I feel like a boiloff plugin has to be very robust if I'm going to take people's fuel away .

On 1/8/2016 at 2:32 PM, Psycho_zs said:

Chelyabinsk does not extend its nozzle when activated.

Really? Did I forget to include BDAnimationModules in the download?

23 hours ago, Psycho_zs said:

..so, boiloff is not linked to temperature or tank volume? I wonder if it is possible to set initial part temperature at launch.

It seems pretty slow. After a week on the launchpad the rocket is still able to go into space, just about 1% of the fuel is gone.

Boiloff won't be linked to temperature until post-1.1 if at all, to wait for more improvements we get to the thermal system. Things are too unstable at high warp. As to volume, it is percentage-based so it's always linked to volume. It will take about the same amount of time for a tiny tank to boil off 100% of its fuel as it will for the biggest tank.

15 hours ago, Fraz86 said:

When transitioning into or out of 10000x or 100000x time warp, ZBO tanks often lose a little fuel despite more than adequate solar panels and batteries.

Now this is a bug I do have to look into!

15 hours ago, Fraz86 said:

The basic mechanics of boiloff appear to work as expected. Fuel loss proceeds exactly as predicted for exponential decay with a rate of 0.025% per hour.

Yay!

15 hours ago, Fraz86 said:

• All ZBO tanks have the same EC requirement (0.25 EC/s). It seems like this requirement ought to scale with the size of the tank.

Yeah that's rather a bug. It should be currently 0.25 EC/s per 1000 u of tank capacity, so it's not being applied properly. a "goal" for this is that the largest 3.75m ZBO tank should need about 10 Ec/s for refrigeration. That probably translates to an effective rate of 0.1 EC/s/1000u.

15 hours ago, Fraz86 said:

As noted in my previous post, 0.025% loss per hour is probably too low to make round trips to Duna using LH2/OX truly awkward. A 7% increase in required fuel mass isn't terribly significant, and thus LH2/OX engines with atmospheric tanks will likely be superior to LF/OX for some Duna missions. Depending on the mass ratio advantage of atmospheric LH2/OX tanks, they might also be superior to ZBO tanks for some Duna missions. I'm afraid these considerations might be confusing/non-transparent to players. I believe the boiloff rate should be at least doubled, in order to make atmospheric more clearly undesirable for interplanetary missions.

Yeah it probably should go up. 

8 hours ago, Captain Sierra said:

It looks like there haven't yet been any changes to LH2 fudge factor or dry mass for atmospheric or ZBO tanks. Do you have some targets in mind?

I do. I thought I had added them (at least I changed the dry mass in the patch, I thought). In my view, there should be a small penalty for LH2 in general for lifer tanks (maybe 10% more dry mass than LFO), and maybe a 20-30% for ZBO tanks. It should be pointed out that adding EC usage to ZBO tanks does effectively increase dry mass as well. 

15 hours ago, Fraz86 said:

The Poseidon Atomic Rocket's description mentions a built-in generator; looks like that hasn't been implemented yet?

*grimaces* I don't know how I want that to function at the moment. I know how I want it to function with NFElectrical installed, because part of the backend for NF engines includes bimodal ntr support. However a simple ModuleGenerator is a bit too simple for this. 

11 hours ago, riocrokite said:

Hey guys, does this boiloff plugin work in the background for not loaded vessels? I'm asking because AFAIR the other boiloff plugin (from RO) affected only one selected vessel

It is commented out in the current build, but should happen in a crude way (catchup on loading that vessel).

8 hours ago, Captain Sierra said:

Re EC: Oh dear. I know this crowd of people. We love complexity and if you give us an inch, we'll take a mile. Give us EC usage and before you know it, people will be asking for full thermal sim. @Nertea you're the one who gets to draw the line on how much feature bloat is acceptable feature bloat.

I stated earlier that EC use was happening.

30 minutes ago, Fraz86 said:

In some situations, it may be important to have a manual on/off toggle for refrigeration on ZBO tanks. For example, a Duna ascent vehicle with cryo engines and ZBO tanks probably won't have enough batteries to keep the tanks powered through Duna's 9 hour nights. It would be desirable to turn off refrigeration at night so the craft doesn't run out of power, and the fuel loss to boiloff would be trivial.

That's probably a good idea to add in. Better to lose some fuel thank control of your craft.

[Starbuckminsterfullerton]

Everything's looking pretty good so far, fuel switch is even working correctly on third party tanks including adding boiloff when appropriate.

I hadn't even seen the Poseidon before now, it looks great! The nozzle is only extending in LOx augmented mode though. Liberator nozzle works, but both engines appear to be missing their exhaust FX.  Engine glow is still there. Test fire was done in atmo, if that matters.

Second on Poseidon generator, if it's a togglable mode and you have implemented it, it isn't showing up. If it's passive/in the background, then it might be working, I just realized that forgot to do a test without launch clamps which of course provide power. 

[Fraz86]

1 hour ago, Nertea said:

Yeah that's rather a bug. It should be currently 0.25 EC/s per 1000 u of tank capacity, so it's not being applied properly. a "goal" for this is that the largest 3.75m ZBO tank should need about 10 Ec/s for refrigeration. That probably translates to an effective rate of 0.1 EC/s/1000u.

Ah, I see, that sounds good!

Quote

I do. I thought I had added them (at least I changed the dry mass in the patch, I thought). It should be pointed out that adding EC usage to ZBO tanks does effectively increase dry mass as well.

Huh, I'm looking at the fuel switch patch right now, and it looks identical to the one from CryoEngines v0.1.11 (except for the addition of ModuleCryoTank).

Quote

In my view, there should be a small penalty for LH2 in general for lifer tanks (maybe 10% more dry mass than LFO), and maybe a 20-30% for ZBO tanks.

To clarify: you're talking about the LH2/OX configuration relative to LFO? And when you say 10% more dry mass, do you mean 10% more dry mass relative to fuel mass, or just straight up 10% more dry mass? If it's the latter, that will require an abysmal dry mass ratio for LH2 - the fuel mass of the LH2/OX configuration will be significantly less than LFO, and most of the fuel mass will be OX, so ending up with higher dry mass than LFO would require a horrendous ratio for LH2 - about 0.787 kg LH2 per kg dry mass, to be specific. Also, we should probably figure out LH2FudgeFactor/mixOXProportion before deciding on specific dry mass ratio numbers. Thoughts?

Quote

It should be pointed out that adding EC usage to ZBO tanks does effectively increase dry mass as well. 

Good point. If the EC requirements are high enough, perhaps it's not even worth the potential non-transparency of giving them higher dry mass than atmospheric tanks? You could give them higher cost, lower maxTemp, and lower crashTolerance to compensate?

Edited January 10, 2016 by Fraz86

[Nertea]

By 10% I mean that if a LFO tanks has for example 10t of fuel and 1t of dry mass, a LH2O tank might have something like 1.1t dry mass.

I'm pretty happy with the Lh2 volumetric ratio right now. I don't see much need in changing it with the proposed changes to dry mass.

And the electrical requirements should probably not be enough to add a huge amount of dry mass. Just some.

[Fraz86]

7 hours ago, Nertea said:

By 10% I mean that if a LFO tanks has for example 10t of fuel and 1t of dry mass, a LH2O tank might have something like 1.1t dry mass.

Based on current volumetric ratios, a tank that holds 10t of LFO will only hold 5.14t of LH2O, of which 4.5t is OX and 0.64t is LH2. The 4.5t OX has a corresponding dry mass of 0.56t (based on the established dry mass ratio of 8:1 for stock LFO & LF tanks). Therefore, in order for the tank to have a total dry mass of 1.1t, the 0.64t of LH2 needs to have a corresponding dry mass of 0.54t, which is a lot. Is that what you had in mind?

Edit:

In fact, it's worse than the above. An LFO tank with 10t of fuel would actually have 1.25t of dry mass. If you want the LH2O configuration to have 10% greater mass, that would be 1.375t. Therefore, the 0.64t of LH2 would need to have a corresponding dry mass 0.815t.

Edit 2:

If you do indeed want dry mass ratios as described above, massPerUnitLH2 needs to be around 0.00009 for stock tanks and 0.00011 for ZBO. Also, massPerUnitOX should be set to 0.000625 to match stock tanks (the OX portion of LH2O ZBO tanks doesn't require refrigeration, and therefore ought to have the same mass ratio as non-ZBO tanks). Here's my excel sheet if you'd like to play with the numbers:

https://www.dropbox.com/s/r3ks6d367umfb9x/Fuel%20swithcing.xlsx?dl=0

Edit 3: Made the spreadsheet more readable.

Quote

I'm pretty happy with the Lh2 volumetric ratio right now. I don't see much need in changing it with the proposed changes to dry mass.

And the electrical requirements should probably not be enough to add a huge amount of dry mass. Just some

Fair enough!

Edited January 10, 2016 by Fraz86

[Psycho_zs]

4 hours ago, Nertea said:

Really? Did I forget to include BDAnimationModules in the download?

Oops, my approach at installing this pack was way too picky.

[Nertea]

@Fraz86, maybe let me try another approach to communicate what I mean:

In a hypothetical scenario with an LFO engine and a LH2O with identical Isp, with the same fuel mass, I want the LH2O setup to perform 10% worse with lifting tanks. How do I achieve that?

[Jimbodiah]

Add mass.

[Nertea]

*dryly* I'm not quite that dim. 

[Jimbodiah]

?