[1.0.2] NovaPunch 2.09. - May 6th - 1.0 Compatibility Update

[Tiberion]

I can beat new pics.

Freyja Crew Service Vehicle (3.75m) Test version 1:

Edit: Link removed, see later posts for newer tests

I only did a minimal amount of balancing (basically Odin numbers X 1.5) but I wanted to get it out there and find the technical issues before I do any heavy tweaking.

No custom IVA, no built-in parachutes (either the stock or the NP radials in 2x or 4x work nicely though)

the top piece has docking ports built in (3 sizes) though its not going to match up visually very well with anything but itself. Might make an attachable docking port to match it. but technically it can dock with any of the clamp-o-trons.

Pod has built-in decoupler to drop the trunk when re-entering.

Anyway, report bugs and issues.

Edited April 6, 2014 by Tiberion

[biohazard15]

Here you go: my first Freyja flight.

Javascript is disabled. View full album

Issues so far:

1) Docking module is named "Odin OTV Flight Pack", and has description of said pack

2) Docking module again: there is 3 "Control from here" buttons in right-click menu. I guess this is due to its compatibility with any port

3) Pod sinks on splashdown (see the last screenshot)

Other things - like glitchy IVA, no portraits for other three crew members, "invisible magic ladders" - these are just visual bugs due to Mk1-2 interior (except ladder - you just need to add some handles to the door to "fix" it)

[Gryphorim]

My only bugbear with this design, and it's a minor one, is the sharp edge at edge of heatshield. Most other pods have some sort of bevel at the edge to smooth out the design. Like I said, minor. Service module is pretty neat and docking port with integrated rcs. 6 man crew? any thoughts on seating placement?

[biohazard15]

Okay, more on the SAS-radiator problem.

It looks like that 2.5m are affected too; tests with 100-ton rocket showed that it too causes inclining.

Funny thing is: stock 2.5m SAS with it's 0.2t weight and size 1 nodes causes NO problems. At all. The same is with stock 3.75m decoupler, which weighs 0.8t - less than your 3.75m SAS.

I have a weird idea here: your SAS are actually resizes of one model - 1.25m SAS. So the question is: how does Unity handle the joints in this case? Does Unity stiff them according to node size (which it should do, according to common logic)? Or does Unity takes this info from original model and just ignores new settings (may be a bug in Unity, actually)?

[Tiberion]

Nice screenshots

The parachute and floating thing was an easy fix, I dropped the CoM down about 80cm in the pod, and it works right now. I'm going to add some handgrips around the hatch and tweak the ladder collider.

That was a good theory about the scale of the model, but I don't think thats it. I just exported the 3.75m model in 1:1 scale and it is still tilting over.

edit: link removed

Try it out and see if you can figure anything else out. If not, I'll have to compensate with some mass bumps, should be alright in the end though.

Edited April 6, 2014 by Tiberion

[CaptainKipard]

Is it me or are the stack nodes not sized correctly? Should there be seven numbers, with the last one defining the connection size?

[Tiberion]

Yeah, the test version from a couple of pages ago updated them as needed with new node sizes. Is there a specific part you think I missed?

[CaptainKipard]

Yeah, the test version from a couple of pages ago updated them as needed with new node sizes. Is there a specific part you think I missed?

Nevermind I downloaded from the first post.

EDIT:

5m parts should have size 4 not 3, no?

Edited April 5, 2014 by Cpt. Kipard

[Tiberion]

5m parts should have size 4 not 3, no?

Size 4 does not exist.

[CaptainKipard]

Size 4 does not exist.

Are you saying that we're limited to the other sizes?

I thought we could make those numbers arbitrarily large.

[NathanKell]

Any size node exists (at least up to 20! never had a stage wider than that); however, AFAIK in the new 0.23.5 joint handling, size 4+ gets no special treatment.

[Tiberion]

You can make any size of part you like. Before 0.23.5 there was only size 0, 1 (default) and 2 nodes (which are represented by the different sized green or purple orbs that snap together. Those sizes exist because the game had stock parts of that size. They added the new size of parts and thus added size3 nodes to the stack node code. They did not add any further sizes.

The size simply affects how far from the "center" that each joint of a stack node (there are 3 of them, think of a triangle imposed over a circle) is placed. The bigger the part, the farther spaced the joints should be to add "leverage" for the physics. Since size3 is the max, the joints don't reach the end of a 5m part, so there may not be enough leverage to remove all wobble for massive parts now.

Nathan: err? since when have bigger nodes even been possible? I've made typos before and the node breaks completely unless it was 0-2 (now 3 I guess)

[CaptainKipard]

Docking port sizes are arbitrary. "size2" is just a name. It could be "mashedpotato" an it would still work with other "mashedpotato" ports.

Any size node exists (at least up to 20! never had a stage wider than that); however, AFAIK in the new 0.23.5 joint handling, size 4+ gets no special treatment.

It's still a good idea to make things forward compatible in case things change in the future.

[Tiberion]

It has nothing to do with docking ports, they just happen to use the same name for sanity.

You can make a docking port that uses size3 as its node name, and use size0 stack nodes on the part anyway.

If I set the 5m parts to 'size4' (which means a 4 at the node_stack config lines) then they would not receive ANY benefit from the new node fixes, they would be super wobbly again (as opposed to wobbly in a few cases now)

[CaptainKipard]

I thought you were talking about docking ports, since "size1", "size2", etc are nodeTypes only used in stock docking ports to identify docking ports that can dock together.

If you say there's no benefit from larger numbers then that's ok. At least I learned something new.

[planeguy868]

I love the new capsule, Tiberion! It looks and flies like a dream.

Now I have 5 different capsules I use in my current save... I probably should cut down.

However, I do have one quarrel with it. There aren't any station parts that are 3.75m, so when I dock the capsule to stations it tends to look a bit silly. I'm not entirely familliar with how difficult it is to rescale models, but would it be possible to release a 2.5m version with a 1.25m endcap? I don't want to burden you with a lot of work for one person's request, so if it's really time consuming, don't bother.

[NathanKell]

Nathan: err? since when have bigger nodes even been possible? I've made typos before and the node breaks completely unless it was 0-2 (now 3 I guess)

Since .20 at least? NovaPunch size 4 (5m parts) have always been fine for me with size 4 nodes (which I have edited them to have since I started playing KSP last May, so as to have compatibility with FAR).

Node size is just an int.

Heck, check out the various part rescales in Realism Overhaul, which often have size 4+ nodes.

[planeguy868]

Actually, I may have found a bug with the capsule's docking port. It doesn't seem to work, and after I tried to use it, it seemed to break all docking port functionality with the rest of the craft I had in-game. I tried editing the .cfg to only be a size 1 port (1.25m), but it still doesn't work. It specifically happened after I installed the Freyja capsule as well.

[KvickFlygarn87]

Is it working in ARM?

[biohazard15]

More on SAS...

Ferram4 suggested that it may be related to part size - but that doesn't seems to be the case, according to decoupler tests.

Test report:

- Experimental 3.75m SAS: It actually got worse, rocket crashes almost instantly. Adding weight helps, but not much

- Stock 2.5m SAS: Rocket is somewhat unstable, it can be corrected easily by turning SAS on

- Stock 3.75m decoupler - 0.83t (engine and fuel tank were connected by fuel line): Rocket is stable.

- NP 3.75m decoupler - 0.8t: Rocket is stable.

- 3.75m SAS with the model of stock 3.75m decoupler - 0.83t: Rocket is unstable.

- 3.75m "SAS-decoupler" with node size definition removed from part.cfg (Node size doesn't defined for stock SAS): Rocket is stable.

- Original NP 3.75m SAS with node size definition removed from part.cfg: Rocket is perfectly stable.

To summarize:

1) http://youtu.be/-24oncXkpQg

2) There is definitely some bug which involves SAS and new joints. Or maybe it's just street magic.

3) To fix this bug, you should remove node size definitions from all SAS modules. Adding some weight also might help, especially in cases like mine, where said SAS module is placed between 5000 kN engine and 300-ton rocket.

Edited April 6, 2014 by biohazard15

[wasmic]

I love the new capsule, Tiberion! It looks and flies like a dream.

Now I have 5 different capsules I use in my current save... I probably should cut down.

However, I do have one quarrel with it. There aren't any station parts that are 3.75m, so when I dock the capsule to stations it tends to look a bit silly. I'm not entirely familliar with how difficult it is to rescale models, but would it be possible to release a 2.5m version with a 1.25m endcap? I don't want to burden you with a lot of work for one person's request, so if it's really time consuming, don't bother.

Only 5? You should see mine. These are just the ones I use for crew transfers:

Mark 1-2 command pod, 5 different vehicles.

NP Odin, 1 vehicle.

KOSMOS VA, 1 vehicle.

Taurus HCV, 1 vehicle.

MrTheBull's PPTS, 1 vehicle.

KSOS, 1 vehicle.

FASA Gemini, 3 vehicles.

FASA Rescue Gemini, 2 vehicles.

cBBp Dragon, 1 vehicle.

Bobcat's Soyuz, 1 vehicle.

Bobcat's Kliper, 1 vehicle.

Bobcat's Buran, 1 vehicle.

And next time I start up KSP, I'll also have the Freyja.

[planeguy868]

Only 5? You should see mine. These are just the ones I use for crew transfers:

Mark 1-2 command pod, 5 different vehicles.

NP Odin, 1 vehicle.

KOSMOS VA, 1 vehicle.

Taurus HCV, 1 vehicle.

MrTheBull's PPTS, 1 vehicle.

KSOS, 1 vehicle.

FASA Gemini, 3 vehicles.

FASA Rescue Gemini, 2 vehicles.

cBBp Dragon, 1 vehicle.

Bobcat's Soyuz, 1 vehicle.

Bobcat's Kliper, 1 vehicle.

Bobcat's Buran, 1 vehicle.

And next time I start up KSP, I'll also have the Freyja.

Heads up, the current Freyja's docking port causes a ton of issues. Methinks Tiberion will release a fix soon, but don't use the docking port. Don't.

I use 6 at this moment, the PPTS capsule from the Angara Pack, a custom resized Orion capsule (to be 2.5m) with an ATV service module, an ESA Minotaur replica (basically a Dragon with a new service/propulsion module from StretchyTanks), the Freyja with a stock docking port, and a custom built Apollo Block 3-5 series (Blocks 3+ and 4 have Soyuz style mission modules, all of them have cut down service modules). Plus the Tiberdyne shuttle.

[biohazard15]

Looks like we found the primary cause of this weird SAS problem - you can't define node size in config. Why is this happens is beyond my mind and imagination - it just happens, let's stop at this point.

During my tests, I have noticed two things:

- First, SAS could really benefit from added weight - it can greatly increase stability under new joint system.

- Second, and most important: SAS rings from NovaPunch are underpowered. For example: we've got stock Inline Reaction Wheel that has weight of 0.3, torque of 20 and energy consumption (ECon for short) of 0.3. Compare this to MPM-125, which stands at weight of 0.27, torque 5 and ECon of 0.5. A question arises: why use MPM-125, except of good looks?

What I propose is: Turn NovaPunch SAS modules into "high cost, high reward" parts. The following is results of my testing and some thoughts.

Format: Weight\Torque\ECon

1. MPM-125

Current: 0.27\5\0.5

Proposed: 0.64\25\1.0

As you can see, I buffed it considerably - but it comes at a cost of slightly increased weight and a big increase of ECon. Heavy 1.25m ships could benefit from it - they won't notice slight weight increase, and they usually have some solar panels or RTGs to generate enough electricity.

2. MPM-250

Current: 0.54\10\1.0

Proposed: 1.12\30\5.0

I'm not completely sure on that one - because you can use 2 stock 2.5m SAS to gain more torque at lesser cost. But people who doesn't like high amount of parts or just want their ships look good (and NP SAS rings looks cool, no doubt here) will use that part. It can be especially useful on space stations - they usually have huge amounts of electricity, and people tends to limit their part count because of possibly lag.

3. MPM-375

Current: 0.83\20\2.0

Proposed: 2.4\45\10

Big and heavy SAS, for a big and heavy rocket, or for a big and heavy station. Will be quite useful for asteroid interceptors.

4. MPM-500

Current: 1.10\40\4

Proposed: 4.8\65\15

Really huge SAS. You need to think about your electricity supply. Useful for huge vessels - E-class interceptors, heavy refuelers, interplanetary motherships.

Please note that these numbers are my estimate; they may be more suited to me and my style of gameplay (big rockets and big interplanetary ships), and not to the others, so feel free to comment and correct.

Edited April 6, 2014 by biohazard15

[Tiberion]

Thanks for the investigation biohazard, I'll make those adjustments. I've made some other engine tweaks as well, and a bunch fo Freyja fixes. I'll hopefully have a new version of all of it together to test sometime today.

They current Freyja test does indeed have broken docking (fixed) and is still experimental, so use it in a cherised save file at your own risk.

[planeguy868]

Thanks for the investigation biohazard, I'll make those adjustments. I've made some other engine tweaks as well, and a bunch fo Freyja fixes. I'll hopefully have a new version of all of it together to test sometime today.

They current Freyja test does indeed have broken docking (fixed) and is still experimental, so use it in a cherised save file at your own risk.

If you restart the game and don't use the Freyja port, it works fine. I'm just using the stock port in the meantime, and moved the top node a bit and removed docking port function so I can mount a clamp-o-tron on top of the Freyja port.

Has the DL link been updated to have the fixed port?