FMP – E2
TITLE OF UNIT: Final Major Project
Email – l.howell0120231@arts.ac.uk
E1 – 20/11/2025
(To come back to the top, simple click on one of the headers!)
Contents:





To create the character’s jacket, I used a blazer frommy previous project. This I resized to the model’s proportions and upgraded to have shoulder pads.



To construct a belt, I simply stitched 4 pieces of fabric together. [see below] This I could sculpt on top of later, to add detail, but fornow, was all I needed.
Then I cut the sleeves where the arm detatches from itself.


I created a bare bones blockout of the mechanical feet from what I imported from Zbrush.
This I created by making one finger, then duplicating it to create the rest.
This formed the basis of what the joint on top would become


Originally, the concept for the joint was more sleek and contained. However, I thought that by having a revolute joint, instead of the spherical one would look better and fi the narrative. Afterall, the character is meant to have DIYed the whole thing, so something bigger and bulkier makes more sense for storytelling.




I also used pistons to extend and contract from the joints as the angle changes. This will later be constrained together, so that when rigged, they will behave like actual pistons

Throughout the modelling process, I have kept a version log. This enables me to track changes, and find it easier to find the correct version I’m looking for.


Next, I tackled the armature of the feet, and the knee rest.
This had several variations. I initially started out by taking a cylinder and extruding parts, but could never find the correct curve.
Eventually, I landed upon using the torus,reducing subdivisions, and using the outer faces as my starting point. I’d first get the curve I was looking for, and then extrude for thickness.
What really helped me – having realised it after one or two iterations – was using instances to maintain the world orientation, but applying it to the positioned mesh. This meant I could extrude and manipulate the object cleanly, whilst it being positioned in real-time, saving small adjustments down the line.
What I struggled with was modelling the joint. Topology really elluded me around the socket. I was modelling a spherical omnidirectional joint, perfect for handling all the necessary angles of the knee, but this meant that the topology needed to be smoother.








Back Robot Hand Modelling Process

I began the “retopology” (I put that in quotes because I’m using it as a basis to start the actual modeling, rather than it being a low poly of a higher)




Finger Caps:


Import to Maya to begin retopology (again)



I didn’t like how many subdivisions it had so…. take 3!!



Used the model to live extrude from, became so wonky, so I did it again (take a shot how many times I redo something)



Finger caps oh how I hated you.





Okay tell me why this took me so many damn tries….






So we started with SVG -> Geometry, then Texture -> Geometry. I genuinely started doing cleanup, then realised how many vertices there were… then I realised that Zbrush would BE SO MUCH BETTER smh
Following this tutorial : Hexagon Grid ZBrush 2018 – YouTube I managed to create a hexagonal grid.


Using a cylinder that had been reduced to a hexagon, I used the preset from the video to create this hexagonal grid. This could then be adjusted to my needs.
This was going really well, and I managed to create a tube using it.

However, I found issue with the fact that the ends just wouldn’t meet. (I figured out once I rubbed my two braincells together)
So, thinking there was no saving this, I started on another method.


I used the ring plugin to create a circular path that I could put in an array.
However, I didn’t realise that they would not manage to join up, so this became ridiculously complicated.

I had found a way to flatten the faces, but the interior was still messed up, and it wasn’t a complete fix.
Fortunately, my two braincells came in clutch, and I realised that I had to mask out the edges so that they’d fit together once joined.

This meant that I could create the tube that I needed.


This looked great when connected to the arm! (Next would be trialling how well it stretched and squished.

References:





A lot of designs use a revolute joint, constrained to one axis, with one singular rotational pivot at the base. This makes it quite nice to rig.


I modelled in sections, for the middle segment, and then mirrored them to create two identical bars. This I think I will make look better for now. But as they say, it needs to exist first.



Leave a Reply