T1 and T2 Endoskeleton Research Thread

Finished the steps and merged it with the main model... so now I'm cleaning up the topology before working on the nut part of the sculpt, separating it and adding the inside screw detail. I'll try and see if I can finish the clean-up tomorrow... no promises as it is tedious.

ShoulderCableConnector-003.jpg


EDIT: Just noticed an error. The threads aren't wide enough. I'll see if I can fix that by simply widening that section... otherwise I'll have to scale it properly before merging and cutting it... Bummer.
 
Thanks.

Separated the two pieces, so now all I need to do is create that lip on the angled part that goes into the nut, as well as create the threads inside the nut, and then finally clean up the topology and then I'm done and can upload it along with the other small bits I made.

ShoulderCableConnector-005.jpg
 
Last edited:
Question: am I overworking this? It's such a small piece and most of what I'm doing now will be hidden anyway. Maybe I should just keep it as a single piece and as simple as the LFS casting!?
 
Question: am I overworking this? It's such a small piece and most of what I'm doing now will be hidden anyway. Maybe I should just keep it as a single piece and as simple as the LFS casting!?
Depends; some will put every single little details that will not be seen by the viewers (only you know it's there);)
Or some, will just put every things together in a simple, but effective, way without getting too crazy with details the viewers will not see.
Your call(y)
 
I simplified it. Made it into a single model again. No threads inside the nut. Just an open hole. Couldn't guarantee that the threads matched the threads on actual hoses... so it would be a lot of work for no reason. So now I just need to clean up the geometry and it is done.
 
So, there’s already a really good 3D scanned endoskull that I would honestly recommend as a base for 90% of people looking to print and finish a pretty good looking endoskull. Especially if your goal isn’t quite as nuts as what we’re talking about here in this thread. It was created by someone who I think was a member here at one time (deltadesignrus). I don’t know them at all personally, I bought their model on CGTrader a few years ago and decided to print it:

View attachment 1693781

(If this link is not allowed please let me know)

I believe they 3D scanned a Thai recast of the M1 kit, then they said they used “photogrammetry” with still frames from T1 to fix the overall head shape and shrinkage issues. Then they tried to add some more accuracy to areas like the ports, etc.

I’m showing this for a few reasons:
  1. I think they did a pretty great job and while it’s not perfect and has issues, it’s the best looking endoskull 3D printable file I’ve seen available online for a decent price.
  2. I think the upper skull and jaw could be a great base for people looking to take it further with modifications. Other parts included are “ok”
  3. If you want a scan that you can print that gets better than this (that extra 10%), I think it would need tons of work on cleaning up and correcting the scan some more. And it would need to be printed it in resin to justify the tedious labor.
I feel there is a problem with this path. Not only in how much work it takes, but that you also cannot help but make choices or take creative license with how you interpret the scanned mesh as you try to move things around and tighten it up or add back detail. I know because I was doing it for hours on end in Zbrush before I abandoned that idea:

View attachment 1693783
Image above shows areas like the upper part of the skull above the brow line smoothed and cleaned up.

So that may still be a path for someone, but I just wanted to put my thoughts on the matter out there.

-Dana
Hello. I'm still on this forum, just not visiting often. Regarding my T1 scan, it is indeed made from a Thai copy of the M1, as you described. However, at the moment, I can't sell it, even though I've invested many years of time, effort, and love into it. Due to being in Russia, I currently can't sell models on most platforms due to sanctions from the European Union and the USA. I also can't receive money from outside my country. I would gladly sell Terminator skull models, but so far, I haven't found a way to do that. Currently, I've found the necessary version of the Tamiya Sand Scorcher gearbox on eBay and am working on modeling these two parts in CAD. In the future, I plan to buy the necessary parts to complete the most accurate model of the Terminator T1 skull. In the description of my profile, you can find a link to Telegram, where I share (in Russian) my observations about the specific features of the structure of the first Terminator skull.
 
Yes they used the casing halves from a Tamiya R/C car gearbox. Here’s a set I acquired that may be the same. I also have the kit bashes detail parts that were used. Took a long time (and to find and acquire some of these bits)
View attachment 1694645

-Dana
No, this is not the version of the gearbox. What you have in the photo is the 2010 version. Before that, there were at least 3 other versions with differences in details and slightly in size. One thing I can say for sure is that parts from the 2010 version could not have been used in the 1984 film.
I also bought the 2010 version and even created an exact CAD model for it (Free CAD Designs, Files & 3D Models | The GrabCAD Community Library). However, it turned out that it has different details and dimensions. Only later did I manage to find the needed mk2 gearbox, but I haven't reversed it into an accurate 3D model yet; I've only done simple scans.
 

Attachments

  • photo_2024-01-14_05-15-50.jpg
    photo_2024-01-14_05-15-50.jpg
    228.4 KB · Views: 68
  • photo_2024-01-14_05-15-52.jpg
    photo_2024-01-14_05-15-52.jpg
    222.5 KB · Views: 70
  • Screenshot 2024-01-16 012107.png
    Screenshot 2024-01-16 012107.png
    592.4 KB · Views: 71
  • Screenshot 2024-01-16 012210.png
    Screenshot 2024-01-16 012210.png
    926.8 KB · Views: 80
Photogrammetry is almost the same as 3D scanning.
Mirror surfaces cannot be scanned, so a water-based color is applied to the surface.
Noise is inevitably generated, so ZBrush is used to polish it.

View attachment 1737922
It's nice to see that people have understood how to use photogrammetry with paint for scanning reflective and transparent surfaces (since most don't take photogrammetry seriously, and it's a shame). In 2017, I also used photogrammetry to scan various objects by simply covering them with a layer of gouache, spraying it with a toothbrush. This creates a slight roughness on the model, but it's insignificant, and the paint speckles allow capturing 100% of the surface shape due to a specific texture pattern on the surface.
 

Attachments

  • IMG_0351.jpg
    IMG_0351.jpg
    12.6 MB · Views: 51
  • 457 (6).jpg
    457 (6).jpg
    4.1 MB · Views: 44
I couldn't get the straight groove option to work. So with this people can just file the edges down... or I may try to alter it.
Have you tried using CAD for modeling (Siemens NX, Solidworks, Fusion 360, Kompas 3D)? In CAD, it's much easier to create non-organic shapes, and the model is parametric, so you can change dimensions at any time, including the dimensions of grooves, lens diameter, and everything else. I've started doing something similar with an eye to check if the microscope diaphragm I bought on AliExpress fits. It does.
 

Attachments

  • Screenshot 2024-01-16 024341.png
    Screenshot 2024-01-16 024341.png
    391.7 KB · Views: 47
  • Screenshot 2024-01-16 024404.png
    Screenshot 2024-01-16 024404.png
    206.3 KB · Views: 39
  • Screenshot 2024-01-16 024424.png
    Screenshot 2024-01-16 024424.png
    613.2 KB · Views: 41
  • Screenshot 2024-01-16 024441.png
    Screenshot 2024-01-16 024441.png
    309.5 KB · Views: 43
Well, I attempted a square cut-out. Actually managed to do it, and in less time it would have taken to clean up the other model. The geometry on the new model has been cleaned up and everything looks good.

View attachment 1757189

Another issue with your models is that they are not suitable for photopolymer printing due to the low polygon count. These polygons will be visible on the printed model!
 
And here's the M1 scan it was made from.

No, you misunderstood. My skull T1 is not made from my second skull T2! Both of my skulls are assembled from different scans of models (mostly M1 and Blown-Up Endoskeleton Skull Fragment), brought to the necessary shape using rather complex (time-consuming) methods. T1 is adjusted to the correct shape and proportions based on screenshots from T1 and official photos from T1. And the T2 skull is customized to fit the form of the SWS Gift skull Basset (an example in the photo). If necessary, I can describe in detail my method so that you can see how useful it is in recreating the original proportions.
 

Attachments

  • gift1a.jpg
    gift1a.jpg
    2.5 MB · Views: 48
Last edited:
I was able to get a REALLY good look at that Propstore T1 Explosion Endoskull when I went to Comicon in 2018. I photographed it from every angle I could (it was behind glass). But, I was able to get enough photos to run it through photogrammetry and get a good enough mesh to reconstruct it in 3D myself in Fusion 360. So, I'm nearly done with what I think is the most screen-accurate depiction of that part anyone has made to date. Others have tried, but I'm able to now see things that nobody gets right. It's actually a very unique part.
After I read it, my hands started shaking from excitement. Will you share something about this back cover with us? I wouldn't mind having those photos to use photogrammetry myself and finally see what I couldn't see anywhere despite searching for 10 years...
 
Last edited:
No, you misunderstood. My skull T1 is not made from my second skull T2! Both of my skulls are assembled from different scans of models (mostly M1 and Blown-Up Endoskeleton Skull Fragment), brought to the necessary shape using rather complex (time-consuming) methods. T1 is adjusted to the correct shape and proportions based on screenshots from T1 and official photos from T1. And the T2 skull is customized to fit the form of the SWS Gift skull Basset (an example in the photo). If necessary, I can describe in detail my method so that you can see how useful it is in recreating the original proportions.
In general, I'll describe the process in simple terms. To achieve correct proportions and details from photos, I used Agisoft Photoscan (now renamed to Metashape). It allows manually placing markers on photos, starting with one marker on all photos, then a second marker on all photos. We place as many markers as possible manually. After that, we press the Calibrate button, and the program calculates the position of virtual cameras in 3D space, replicating the original camera positions. Simultaneously, all markers in 3D space align precisely where the details on the skull were (where we placed the markers). This is done with considerable accuracy. Next, we export the markers and cameras to 3DsMax or Blender, where we can set an appropriate screenshot as a background for each camera.

Then, based on the camera backgrounds, we draw splines (3D curves). For example, we draw a spline for the forehead seam and bend it to match the seam on all cameras (screenshots). As a result, we obtain a set of three-dimensional lines that precisely replicate the lines of the skull (fortunately, the Terminator has many lines on the skull). After that, having these lines, we adjust, for example, a 3D scan to them. This way, the scan takes on the proportions of the original. We cannot control the curvature of some surfaces using this method, which heavily depends on the quality of the original model we use for scanning. However, all seams, screws, holes, etc., fit very precisely, along with all size relationships and proportions.

Below is my example obtained using this method. My scans are adjusted based on these splines.

In this video, as an example, another program is used for the same purpose, but I didn't use it.
ImageModeller example -
 
Last edited:
Back
Top