Rescale Factor Need Help!

[AR3S_TGL]

Hello y'all! I am starting part creation and I have found my first thing (of many) that I don't understand.

I think that the rescaleFactor is like tweakscale

Is this right? If not, please tell me what it is.

[MacLuky]

Well the name gives it away ;-) it's the factor by which the model is scaled. Typically just leave it at 1 but if you want your part to be bigger you can go for 2. This does mean that the attachment points need to move by the same factor, and for fairness probably mass and cost etc too.

[InsaneDruid]

It is one of rather many points in the cfg that can be used for scaling. But is is not the best way as there are (at least there where) issues when using a rescalefactor of >1 like changing sizes when resetting a flight back to start and such things. In this context, please note that there are multiple ways of defining meshes used in a ksp model and also multiple ways of defining nodes used to connect KSP models.

The old way of defining a mesh that is used by a models from the associated config was:

mesh = model.mu
scale = 1
rescaleFactor = 1

the mesh has to be located in the same directory as the config is.

A better way to do it is:

MODEL
{
    model = HGA/Proton/parts/stage1
    scale = 0.60976,0.60976,0.60976
}
rescaleFactor = 1.0
scale = 1.0

as you can see, there is now a full path associated with the mesh (the mesh's name in the example is stage1.mu) and there are also three independent scaled axis (x, y, z) for this mesh. In this example, I used the rather crude looking scale of 0.60976 as the parts used here are modelled to their original size (4100mm diameter) in a metric environment in blender and scaled to be 2.5 meter parts for stock KSP (2500mm/4100mm=0.60976). This allows for easy modelling (as the original values for every measurement can be used) and also for easy exporting to stock KSP or upscaled mods like realism overhaul where 1:1 models are used instead of the 64%-ish scale of stock KSP. The scale and rescale values of 1 outside the MODEL{} element can be kept at 1 to avoid the problems described above that arise when using anything but "1". Using the MODEL{} declaration also allows the use of more than one mesh in more than one location to be used as one KSP model.

 

For the nodes that connect the models the following notation can be used:

node_stack_bottom = 0.0, -0.47924, 0.0, 0.0, -1.0, 0.0, 2
node_stack_top = 0.0, 1.19319, 0.0, 0.0, 1.0, 0.0, 1

or

NODE
{
    name = top
    transform = stage1_nodeTop
    size = 3
    method = FIXED_JOINT
}

NODE
{
    name = bottom
    transform = stage1_nodeBottom
    size = 3
    method = FIXED_JOINT
}

The first one uses coordinates that have to be manually extracted while modelling and also have to be manually scaled when using a bigger or smaller mesh like described above. They are especially awkward to use when radial attachment points are used that don't align perfectly with one of the major axis of x, y or z of the model. Also they are not perfectly reliable and a bit inaccurate due to rounding problems.

The latter notation might seem intimidating and overly complicated at first, but once learned it is easy to use and completely eliminated the need for any scaling and manual extraction of coordinates. Simply put an empty transform on the model on exactly the location the node shall sit, rotate in in the desired direction and export the model like you normally do.

Then each node is defined as:

NODE
{
    name = <name of the node, like "top" or "bottom" or "engine1"
    transform =<name of the empty transform on the model>
    size = <size of the node from 0 to 3>
    method = FIXED_JOINT
}

See this thread for further reference:

You can see, by using the more robust MODEL{} and NODE{} declarations you can get an easily and reliably scaleable mesh with no side issues and perfect precision.

I learned them for my first model and used them on everything since then with great success.

 

Edited May 5, 2017 by InsaneDruid