[1.4.*] [2.5.3] (2018-04-06) UbioZur Welding Ltd. Continued

[Alewx]

[quote name='aagun123']Want some help:

It says "The hull has more than 255 polygons. This is invalid" when i use welded part in ver 1.0.5, and it makes lags.

Thx in advance[/QUOTE]

Can you give a bit more Information?

[Fwiffo]

Any idea why this weld produces a max temp of 1200, when all the parts are 2000?

Parts are (in order of attachment, starting from root): TT-38K Radial Decoupler, Modular Girder Segment, Probodobodyne QBE, 2x Sepratron I. I checked the stock cfg files for all of the above and do indeed seem to be 2000. Unless ModuleManager or some other one of my mods overwrote that, in any case...

[spoiler=Image][IMG]http://i.imgur.com/AKmfsWH.png[/IMG][/spoiler]

[B]EDIT: [/B]NEVERMIND. My fault. Didn't realize probe core is 1200, not 2000. Sorry for the white noise. Nothing to see here. Move along. Edited November 23, 2015 by Fwiffo
Oops

[aagun123]

[quote name='Alewx']Can you give a bit more Information?[/QUOTE]

It runs perfectly whit stockparts only.

But if a ship uses a welded part, there will be many errors "[COLOR=#333333][I]The hull has more than 255 polygons. This is invalid.[/I][/COLOR]" in KSP logfile, especially when ship warps.

and when these errors occur, the program lags.

Oh my poor english ...;.;

[Fwiffo]

Still toying with this wonderful mod :-). I think I may have found a bug and (hopefully) some info to help you fix it.

When I welded [URL="http://tinyurl.com/WeldmentSrfAttachBug"]this part[/URL] (choosing the TT-38K Radial Decoupler as the partbranch), the resulting weldment's [FONT=courier new]node_attach [/FONT]ended up in the wrong place. It should be near the right edge of the decoupler, but was closer to the weldment's COM. I was able to fix it through some manual work, but I'm hoping it can be addressed in the next update?

I think the mod correctly took into account the 0.01 offset between the original TT-38K's COM and its surface attachment point, but then did not add the delta between the TT-38K's original COM and the new center of mass of the welded part.

[spoiler=Image][IMG]http://i.imgur.com/LggMPoX.jpg[/IMG][/spoiler]

[I](In the ZIP: Test5.craft is the pre-weld, Test5.cfg is the post-weld, Text-5-Fixed.cfg is my manual fixup, Test5-Meas.cfg was used to help me measure the distance between the girder's node and the surface attach node of the decoupler)[/I]

[Alewx]

[quote name='aagun123']It runs perfectly whit stockparts only.

But if a ship uses a welded part, there will be many errors "[COLOR=#333333][I]The hull has more than 255 polygons. This is invalid.[/I][/COLOR]" in KSP logfile, especially when ship warps.

and when these errors occur, the program lags.

Oh my poor english ...;.;[/QUOTE]

[URL="http://forum.kerbalspaceprogram.com/threads/92229-How-To-Get-Support-%28READ-FIRST%29"]Basic rule of thump for assistance by Mod creators.[/URL]
Can you provide more than just this single error message?
Like, parts post/pre welding and log?



[quote name='Fwiffo']Still toying with this wonderful mod :-). I think I may have found a bug and (hopefully) some info to help you fix it.

When I welded [URL="http://tinyurl.com/WeldmentSrfAttachBug"]this part[/URL] (choosing the TT-38K Radial Decoupler as the partbranch), the resulting weldment's [FONT=courier new]node_attach [/FONT]ended up in the wrong place. It should be near the right edge of the decoupler, but was closer to the weldment's COM. I was able to fix it through some manual work, but I'm hoping it can be addressed in the next update?

I think the mod correctly took into account the 0.01 offset between the original TT-38K's COM and its surface attachment point, but then did not add the delta between the TT-38K's original COM and the new center of mass of the welded part.

[spoiler=Image][url]http://i.imgur.com/LggMPoX.jpg[/url][/spoiler]

[I](In the ZIP: Test5.craft is the pre-weld, Test5.cfg is the post-weld, Text-5-Fixed.cfg is my manual fixup, Test5-Meas.cfg was used to help me measure the distance between the girder's node and the surface attach node of the decoupler)[/I][/QUOTE]


well that is not an easy one, while the fix is pretty simple, the workaround in the process is more problematic. How shall the mod know if the part is an attached stack part, radial part, radial attached or even something worse?
How many tried did it take you to get the node_attach the correct position? Edited November 23, 2015 by Alewx

[Fwiffo]

Thanks for replying so quickly.

[quote name='Alewx'][COLOR=#333333]How many tried did it take you to get the node_attach the correct position?[/COLOR][/QUOTE]
Two tries, over several hours. Mostly because I was learning / reverse-engineering the relevant cfg stuff (still new to this), figuring out coordinate system and how my parts (like the girder) were rotating/translating within it, examining stock cfg files, drawing diagrams, etc. I finally figured out the easiest approach (in this case) was to take the the decoupler's model location coordinates within the weldment, and add the offset of it's node_attach (0.01) to that value. Then I got perfect results every time. Also helped to use vector precision 6 instead of 3.

[quote name='Alewx'][COLOR=#333333]How shall the mod know if the part is an attached stack part, radial part, radial attached or even something worse?[/COLOR][/QUOTE]
I don't entirely understand this question, or it exceeds my existing level of KSP knowledge :-). Any chance you could describe a couple hypothetical weldments that would exemplify this ambiguity?
I thought node_attach only comes into play when you're attaching your part to the surface of another one, so not sure why the first one is relevant. I expect you'd always follow the attachment methodology of the "root" partbranch. But I get the sense perhaps I'm missing the meat of your question.

[Alewx]

[quote name='Fwiffo']Thanks for replying so quickly.


Two tries, over several hours. Mostly because I was learning / reverse-engineering the relevant cfg stuff (still new to this), figuring out coordinate system and how my parts (like the girder) were rotating/translating within it, examining stock cfg files, drawing diagrams, etc. I finally figured out the easiest approach (in this case) was to take the the decoupler's model location coordinates within the weldment, and add the offset of it's node_attach (0.01) to that value. Then I got perfect results every time. Also helped to use vector precision 6 instead of 3.[/QUOTE]
That is why precision is made an option to configure on personal desire^^

[quote name='Fwiffo']I don't entirely understand this question, or it exceeds my existing level of KSP knowledge :-). Any chance you could describe a couple hypothetical weldments that would exemplify this ambiguity?
I thought node_attach only comes into play when you're attaching your part to the surface of another one, so not sure why the first one is relevant. I expect you'd always follow the attachment methodology of the "root" partbranch. But I get the sense perhaps I'm missing the meat of your question.[/QUOTE]
The problem is the mod can not know what kind of connection is planned for a weldment, will it be a stacked part, a radial attached part via node, or just a radial part, or even something completely different.
That is the reason the part gets the node_attach to the CoM, just like stock parts handle it.

[Hysterrics]

Oh boy, time to continue work on the ISS without killing my computer.

[Alewx]

[quote name='Sequinox']Oh boy, time to continue work on the ISS without killing my computer.[/QUOTE]

Depends mostly on the computer :D

[Fwiffo]

[quote name='Alewx'][COLOR=#333333]The problem is the mod can not know what kind of connection is planned for a weldment, will it be a stacked part, a radial attached part via node, or just a radial part, or even something completely different.[/COLOR]
[COLOR=#333333]That is the reason the part gets the node_attach to the CoM, just like stock parts handle it.[/COLOR][/QUOTE]
No problem with it getting node_attach, just want that node to be positioned in a sensible place. The other nodes take care of the stacked cases just fine, and AFAIK the weldment's node_attach is [I]only [/I]used when you radially attach the weldment to something.

So I'm not sure the mod really needs to know what you intend with it (although I concur such knowledge could be useful). Would it not make sense to always put the node_attach where the root part's original node_attach was (provided it has one)?

Sorry if I'm missing some subtleties here.

-Fwiffo

ps. By root part, I mean the part you select as the partbranch.
pps. From what I've seen so far, a lot of stock parts do not have node_attach at the COM. I think they tend to be along COM in two axes but usually on (or nearly on) a surface on the third. Edited November 26, 2015 by Fwiffo

[Alewx]

[quote name='Fwiffo']No problem with it getting node_attach, just want that node to be positioned in a sensible place. The other nodes take care of the stacked cases just fine, and AFAIK the weldment's node_attach is [I]only [/I]used when you radially attach the weldment to something.

So I'm not sure the mod really needs to know what you intend with it (although I concur such knowledge could be useful). Would it not make sense to always put the node_attach where the root part's original node_attach was (provided it has one)?

Sorry if I'm missing some subtleties here.

-Fwiffo

ps. By root part, I mean the part you select as the partbranch.
pps. From what I've seen so far, a lot of stock parts do not have node_attach at the COM. I think they tend to be along COM in two axes but usually on (or nearly on) a surface on the third.[/QUOTE]

Hmm strange the tanks normally have them at the CoM, you looked at some radial parts?

[Kershu5]

I fear I am missing something I cannot yet find in the updated forums.  Every time I start welding and "save" my weld, the craft is immediately unloaded and deleted.  It'll say "[WeldingTool]   ----- End Welding ----- | X Parts welded" and then it'll say it's welded a part, suddenly the log says "deleting part name.name" until the entire craft has been deleted.  I see no way to load them, either?

 

e: I should clarify, I don't see where I can load the weldments from even after reloading.

ee: Ignore that, I just saw my mistake.  Derp.

Edited November 30, 2015 by Kershu5

[Fwiffo]

On 11/26/2015, 4:55:08, Alewx said:

Hmm strange the tanks normally have them at the CoM, you looked at some radial parts?

I don't think that's correct.  e.g. From fuelTankJumbo-64.cfg:

node_attach = 1.25, 0.0, 0.0, 1.0, 0.0, 0.0, 1

Although at the COM vertically, the node is on the right edge (it's a 2.5m diameter tank).  i.e. It's on the side surface of the tank, centered vertically.  I checked a couple other tanks at random and didn't see any where node_attach was at 0,0,0.

[Alewx]

5 hours ago, Fwiffo said:

I don't think that's correct.  e.g. From fuelTankJumbo-64.cfg:

node_attach = 1.25, 0.0, 0.0, 1.0, 0.0, 0.0, 1

Although at the COM vertically, the node is on the right edge (it's a 2.5m diameter tank).  i.e. It's on the side surface of the tank, centered vertically.  I checked a couple other tanks at random and didn't see any where node_attach was at 0,0,0.

Ok we can try that.

[BlueCanary]

Is welded boats leaping 4000 metres into the air when they load on water a known issue? I know someone IRL who's having the same problem so I'm 90% sure it's not just something I'm doing wrong, I have a welded aircraft carrier sitting in the sea near KSC and when it loads visually at 2.5km it's fine but at around 100-300 metres when it loads properly it is catapulted thousands of metres up at supersonic speeds. This only happens on water. I'm using 1.0.5 with no other mods except Hyperedit and Pilot Assistant, so I'd be very surprised if they had anything to do with it, but I can try with them removed if you want.

Do you know any ways to fix/avoid this or anything? Also, if there's any logs or anything that'd help I'm happy to send you them.

Edited December 3, 2015 by BlueCanary

[Temeter]

On 3.12.2015, 19:15:28, BlueCanary said:

Is welded boats leaping 4000 metres into the air when they load on water a known issue? I know someone IRL who's having the same problem so I'm 90% sure it's not just something I'm doing wrong, I have a welded aircraft carrier sitting in the sea near KSC and when it loads visually at 2.5km it's fine but at around 100-300 metres when it loads properly it is catapulted thousands of metres up at supersonic speeds. This only happens on water. I'm using 1.0.5 with no other mods except Hyperedit and Pilot Assistant, so I'd be very surprised if they had anything to do with it, but I can try with them removed if you want.

Do you know any ways to fix/avoid this or anything? Also, if there's any logs or anything that'd help I'm happy to send you them.

I have (potentially) similar issues, for me, my ships just insta drop to the ground of the ocean when loading physics after spawning in in. Was always connected to specific ship types, when I reconstructed a new craft using the same wielded parts, they tend to work. It's a bit annoying, but no clue what to do about it.

edit: Thinking about it, I remember reading about vessels sometimes having issues with deciding if they are landed (=ocean floor) or on the water surface, so it might not be connected at all. Still, only encountered the issue with welded ships.

Edited December 7, 2015 by Temeter

[Guest Space Cowboy]

I am continuing to use this fantastic mod on the hopefully soon to be released Millennium Falcon.

Unless I am missing something, there is still a huge problem with weldments - attaching them where you want in the hangar.

In addition, I have a suggestion (not related (?) ): Please make a button to turn off node generation. Erase all nodes except root node. Let the user make nodes with radial attachment point or cubic octagonal strut parts.

 

[ThePixeledFox]

I am making a docking ring with mk2 crew cabin but only the root part lights work and none of the ring will light up , yet i do not have this issue if i make a ring with the mk1 crew cabin is this a bug ?

[Kershu5]

So I tried to make a thing with welding, but when I reloaded the database and tried to launch it, every time I try to right-click it to use its various options, I get "NullReferenceException"s.  It didn't do this with the other parts I experimented with despite most of them being from mods like MKS and such, should I have not used the alt+F12 menu to reload the database after saving the new welded part?  Or is it something else?

 

e: never mind, I figured out what I did wrong, it all works now.  Yey.

Edited December 11, 2015 by Kershu5

[Guest Space Cowboy]

http://forum.kerbalspaceprogram.com/index.php?/topic/127577-millennium-falcon-105-progress-but-on-hold-for-now-video/

I am down to 50 parts but not getting anywhere near the fps I was hoping for. Suggestions?

 

Edited December 11, 2015 by Space Cowboy

[legoclone09]

Love this mod and always have, could you get it on CKAN if/when you get this on KerbalStuff?

[CommanderSpock]

I'm trying to weld a 2,300 part ship, I can select it all but it won't weld. The instructions for saving it don't come up. Is there anything I should do?

[Speadge]

On 11.12.2015 at 6:17 PM, Space Cowboy said:

http://forum.kerbalspaceprogram.com/index.php?/topic/127577-millennium-falcon-105-progress-but-on-hold-for-now-video/

I am down to 50 parts but not getting anywhere near the fps I was hoping for. Suggestions?

i have the same problem with a huge Station.. 37 parts and lagging like 300.

I think its the loading of the models at last that takes the cpu-power, less the part interaction...

And the count of the "models" is the same at all, even with 300 times the same model, it lags as hell if it is weldet to 1 part

Edited December 30, 2015 by Speadge

[Guest Space Cowboy]

23 hours ago, CommanderSpock said:

I'm trying to weld a 2,300 part ship, I can select it all but it won't weld. The instructions for saving it don't come up. Is there anything I should do?

Try just waiting a really long time. Give it 5 minutes even. Welding my falcon takes quite awhile. Give it a few minutes, knowing nothing of your system specs, and reading your symptoms, this would be my recommendation.

[CommanderSpock]

1 hour ago, Space Cowboy said:

Try just waiting a really long time. Give it 5 minutes even. Welding my falcon takes quite awhile. Give it a few minutes, knowing nothing of your system specs, and reading your symptoms, this would be my recommendation.

okay, I'll try that.