Modelling small-scale figures – Part 1: ‘twisted wire’ armature

Please note if you’re reading this older post that this and later articles on figure modelling can now be found in the Methods section under Modelling

Here is a full, step-by-step account of making a simple twisted wire armature suitable for 1:25 scale figures. I’ve been meaning to put this together for some time, since long before WordPress, because in my book I suggest a more involved method soldering brass and there was a need for an easier alternative. A good armature is not merely the support for a soft material which would otherwise lose its shape; it should also provide guidance on where to model, in the way that a version of the human skeleton would when modelling on a larger scale. For this reason I’ve found it valuable to define torso and pelvis mass by adding thin cardboard pieces.

I also felt that I definitely needed to choose a wire that was easy for anyone to find again and, hopefully, always available. It’s not enough to say ‘use a very thin wire’ .. because the right kind of thin wire isn’t easy to find! Luckily I’ve found the perfect one for this purpose, a thin plastic-coated garden wire from B&Q called ‘B&Q Value Plant Twist Tie’ costing £1.98 for a 50metre spool. An almost identical kind of wire and spool can be found sometimes in £shops. The whole armature is made from one continuous length of wire, doubled up and twisted fairly tightly together. Although the plastic ribbon adds to the bulk, you will see from the photo below that it compresses well when twisted. If twisted firmly and evenly the final thickness is only a little more than 1mm, representing 3cm at 1:25 scale. What’s more, the twisted surface gives a much better key for modelling material than smooth wire alone. This form of coated wire is also very strong (although it doesn’t have to be so much for this purpose). For example I’ve been testing it as a possible alternative to thicker aluminium wire for animation puppets and at the last marathon ‘bending session’ I reached 100 times back and forth on a fairly fixed point until I gave up .. it seemed completely unaffected!

wire armature

Here below are the scaled templates I’ve drawn up, used as guides throughout the making process. The first is for a 175-180cm male, the second for a 160-165cm female. I’ve uploaded them at proper size (right-click and save) so they should print out at the right scale but if not they are available to download (along with some other modelling reference sheets) on an older resource site of mine:

These template drawings need a bit of explanation. The proportions are based on an ‘idealised average’, if that makes sense, i.e. male a little taller and with broader shoulders than real average and female a little more shapely. The stick figures include hands and feet, just so that it’s clear where these are .. but I usually recommend snipping the wires for arms and legs just halfway into the hand or foot symbol. That is explained below. The torso and pelvis ‘plates’ are indicated by the lighter outlines. Just about any thin (i.e. up to 1mm) strong card can be used, or even thin wood, but not plastic because this will distort with the heat of baking if Super Sculpey is to be used. The drawings also indicate where elbow and knee joints are likely to come, as reference for later modelling. The ‘high kicking’ legs drawn with a dashed line are just there to be able to measure the full length that the leg wires should be outstretched, before they’re glued to the sides of the pelvis plate.

male soft wire armature 1to25

female soft wire armature 1to25

A 70cm length of wire is needed for a 1:25 scale figure, allowing for some extra in case of mistakes. The wire needs to be folded to find the mid-point and the two strands twisted together from there forming a little loop for the head. The neck is then twisted down to the shoulder line. The head-loop should be kept small, certainly no bigger than the size indicated on the drawing. One problem with achieving realistic figures in this small scale is that the heads invariably end up too big, usually too round .. what I call the ‘Playmobil syndrome’ .. so it’s best to guard against this from the beginning. Although this wire is easy enough to twist using just the fingers it may help to use a cocktail stick as a lever, shown below.

twisting wire

The length of the ‘neck’ should then be checked against the drawing (below I’ve placed it a little underneath just so that you can see the drawing) and the two strands of wire splayed out.

checking shoulder line

Each strand will first form an arm, and this is done by making a loop which returns to the centre. To be safe make the loop extend about 1.5cm longer than the drawn arm, because the wire contracts when twisted.

forming loop for arm

Hold the two strands of wire firmly together at the bottom of the ‘neck’ and start twisting from there. Twisting is much easier, and more regular, if after a few twists you open out the loop to use as a lever.

twisting loop for arm

Keep going until the loop becomes a bit too small to manage with the fingers and then check against the drawing. The arm needs to be long enough to extend just halfway into the hand shape, not right to the fingertips.

twisting loop for arm

At this scale hands are so slight and the wire is too thick for them. This way there’s just enough wire to attach a small ‘flipper’ shape to it, to suggest a hand.

trimming to length

The wire can be cut with an old pair of scissors, but it’s much easier and more accurate to use proper wire snippers below. Actually they’re known as ‘wire nippers’, just for cutting wire, with a head like a puffin. £shops don’t tend to have them but the cheapest one’s I’ve found in London are from the Danish chain ‘Tiger’, small ones for £2 usually found in the crafts section.

wire snippers

Below, of course next comes the other arm done in exactly the same way. But make sure that you do this with the right strand of wire remaining. There are two .. take the longer one! It may seem obvious to some, but anything can happen during a lapse in concentration!

other arm

But here it’s also important to try to make the twist start as level as possible with the first arm.

keeping shoulder line straight

Next, twist the two remaining strands together to form the spine as far down as the top of the pelvis shape.

twisting down to pelvis

This is where the armature starts becoming something more identifiable .. more believable. Use the drawing as a shape template to cut thin pieces of card for the torso and pelvis (see the paragraph at top, just above the template drawings). Tack the torso piece to the drawing using UHU or double-sided tape and position the wire figure over it, pressing down so that the wire lies as flat as possible on the cardboard.

positioning for torso

Below I’m fixing the torso in place with superglue, adding baking soda (bicarbonate of soda) to accelerate and thicken the glue. It’s the quickest and strongest method. Superglue alone can’t be made to stick enough to this kind of plastic, but adding the powder seems to ‘cement’ it solidly.


Above and below, I’ve dripped thin superglue over the wire pieces, enough to run onto the cardboard, then used a plastic brush loaded with the baking soda to drop it onto the glue. It sets almost immediately into a rock-hard mass, filling gaps between the wires and the cardboard.

using baking soda

hardened superglue

Other things could be used for this such as 2-part epoxy glue, which can be fast, but this is instant and the rough surface is a good key for any modelling clay. Using baking soda as an accelerator and gap-filler with superglue has been known about for some time but as yet I’ve found no complete explanation as to why it works so well. I’ve read that superglue, apart from needing a small amount of moisture to set, also needs neutral ph conditions (i.e. not too acidic) and since baking soda is often used to create just that (i.e. used as an antacid, for example) this may have something to do with it. It may also just be physical rather than chemical. When superglue is filled with particles of a certain type the chain molecules that form as it sets don’t have so far to go perhaps, so setting is quicker?

But getting back to this, the next stage is to repeat the process for the legs …


… and cement them to the pelvis piece in the same way. The surplus strands of wire can be cut off after gluing.

gluing to pelvis shape

The legs need to be secured along the top of the pelvis piece first, then bent round the outside edge and glued in place. That’s another great advantage of superglue, to be able to glue in small instalments!

gluing to pelvis

The armature is now ready to be modelled upon! My technique of modelling, whether I’m using Super Sculpey ( a polymer clay which needs heat to harden) or Milliput ( a 2-part epoxy putty which cures on it’s own), is to build up very slight body masses first, just enough to make those parts of the wire inflexible but keeping the joints free so that the figure can still be repositioned. It’s easier to work on the general shapes of these body masses, and to keep the symmetry above all, while the figure is still a ‘spreadeagle’.

But the modelling is a separate chapter, which will follow at some point …

armatures in progress


Making walls – Part 3

In Part 1 I introduced a special method of building a model wall which involves ‘laminating’ a soft foam layer onto a cardboard (or Pvc) cut-out. This makes for a much stronger (also lighter) construction generally, as well as making it easier to work to a required thickness without the problems involved with cutting through other thick materials and keeping the edges precise. The third advantage is that the foam on one side lends itself more readily to heavily textured effects (for a possible exterior) whereas the harder layer on the other side is a suitable base for interiors.

simple interior wall drawing

This Part 3 follows directly on from Part 1 and looks at applying details to the interior side, in this case the simplest version drawn above (Part 2 had jumped ahead a bit, looking at more ambitious effects with Kapa-line foam). Drawing up .. always/however simple the task/without question .. is the first requirement! Finding the right scale of material to convey a convincing look, with an appropriate visual weight or lightness, is the next requirement. Doorway and window frames, skirting board etc. can be built up using a variety of materials .. card, wood or plastic .. as long as they can be cut cleanly and are thin enough for the scale intended. For example here the scale is 1:25, so standard mountboard (at almost 1.5mm thickness) is really far too thick. Mountboard is also too fragile when cut into very thin strips (the edges don’t stay sharp and the top layer is also likely to separate). Much better results can be achieved with either stencil card, 0.8mm obeche wood, 0.5 -1mm styrene sheet or, as I’m using here, 1mm Palight foamed Pvc.

Palight foamed Pvc doesn’t come any thinner than 1mm but it just about works for wall details in 1:25 scale especially if the edges are softened. That’s the other big advantage over card .. that it can be sanded, either to shape it or neaten up edges. It can also be given a convincing woodgrain texture, and that’s what I’m doing first below with a small piece of the Pvc. Here I’m using 120grit sandpaper on a sanding block, which needs to be pressed/dragged firmly along the plastic in one direction. Coarser sandpaper (i.e. up to 40grit) can be used for a larger scale effect.

graining Pvc

The effect doesn’t show up that much until it’s rubbed or washed with acrylic or woodstain, shown by the collection of samples below.

wood simulation

Here I wanted to build up the doorway and window frames in two layers, and position the first layer very slightly outside the edge (rather than right on it) so that there is an ‘extra’ little step visually. To keep the positioning regular I drew guidelines (i.e. the outer limits of the frames) around the openings.

marking frame outline

The art, the science and ‘sense’ of cutting

Below are the strips cut to make the first layer. I’m sure I must have said many times that there’s a whole little book to be written just on cutting with the scalpel! At the very least, anyone not practised in cutting needs to consider it a subject in itself which needs to be rehearsed, explored and ‘made peace with’ as far as possible before being able to do anything else. Scale model-making of this kind is so dependent on being able to cut a straight line in the right place. It sounds so simple .. but it’s not! It can’t just be taken for granted that everyone will be able to do this with just a little practise and often people who could otherwise become excellent makers are put off the whole idea of model-making just because this one aspect is never really ‘conquered’. Here are a few guidelines:

If you’re using a material for the first time you should take a while just to get a sense of how it cuts i.e. starting with how steady the metal ruler will lie on it, how resilient or giving the surface is to the initial pass with the blade, how many passes are needed to cut through cleanly without excessive pressure. If you don’t feel confident that the ruler will stay where you’ve put it, you either need a better ruler or you need to do something so that it will grip better. Flat steel rulers will certainly need a strip of masking tape on the back at the very least but sometimes this isn’t enough so pieces of double-sided tape could be added provided they won’t damage the material.

You should also rehearse what it feels like to run the tip of the blade steadily along the metal edge, without necessarily cutting at all. It should feel locked there, able to run freely along but not to depart from the edge. The scalpel blade is slightly flexible and it should be pressed hard enough into the metal edge so that it flexes just a little.

There are no special prizes for being able to cut through in one go! The first pass with the knife should simply be to establish a guiding ‘scratch-line’ which only has to be deep enough to be found again with the tip of the blade. One’s focus at this stage should be more on the edge of the ruler than the material to be cut. Pressure comes afterwards, once one’s established this line and it shouldn’t matter how many passes it takes to cut through. If you’re having to press so hard to get through the material that you can’t control the straightness of the cut anymore it means one or more of the following:- the material is too tough or thick to be cut with a scalpel and you will have to try with a Stanley knife or failing that a saw; you can turn the sheet over and try cutting in exactly the same place on the other side (when cutting thick materials it’s the friction on the blade that becomes the problem and starting ‘new’ from the other side often works); you need to build up some more strength in your hand and arm through practise.

If, for whatever reason, the ruler moves while cutting, don’t try repositioning it by eye. Put the tip of the scalpel in the beginning of the line you’ve started and slide the ruler up against it. Holding onto that position put the tip of the scalpel in the end of the line and move that end of the ruler against it. You might need to adjust, beginning and end, a few times.

It’s worth asking yourself consciously whether you’re working under the best conditions or whether they can be very simply improved? For example .. is the cutting matt flat and smooth or is it more like a Jackson Pollock? Can much of this be scraped off? Is the cutting matt really flat on the table or are there small bits of scrap under it? Have you really got enough proper light to work by? .. in particular, can you see your marked line clearly enough or is the edge of the ruler casting a shadow over it?

Usually with thin materials (i.e. up to 1mm) the angle of the cut edge, in other words whether it’s at a right-angle to its surface or not, doesn’t matter so much. Generally, if one’s holding the scalpel normally it will be fine. But if over 1mm thick it can matter, especially if the edge is to be glued on something else at a right-angle. If using foamed Pvc or wood it would be normal practice to straighten the cut edge using a sanding block and this will even work with foamboard or some types of cardboard. Even so it’s best if one gets used to holding the scalpel upright in the first place. It’s much easier to maintain the knife upright if you can actually see the angle while cutting, i.e. by cutting the line in the direction straight ahead of you rather than side to side. Especially when cutting longer lines it’s usually better to stand up for this so that you can reach over the work properly and use your own body as a ‘measure of uprightness’.

cutting strips

If you’re using a standard surgical-type scalpel (i.e. Swann-Morton No.3, which I would always recommend) the best blade to use with it is the ’10A’. It’s the most general-purpose but also the most precise. Changing the blade (i.e. when it gets blunt) can be a nightmare with a new scalpel because the fixing is often very tight, making it hard to slide the blade either off or on without fear of injury. The only way to solve this (until it wears down a little with use!) is to file into the blade channels a little. Below is not intended as a solution to this, but it does help to know that blunted blades needn’t always be replaced. They can quite easily be sharpened on a piece of ‘wet and dry’ or Emery paper (usually best 600-800 grit) by stroking the blade firmly at a shallow angle, a few times each side. It’s usually only the very tip of the blade that gets blunt so it’s best to focus on sharpening just this small part, flexing it a little into the paper.

sharpening scalpel

Here .. back to the actual work now! .. I’ve chosen to mitre the pieces of doorway/window frame together as they would be in reality, that is, to cut the joining ends at a 45degree angle (as picture frames are also usually made). In the real-scale world this is done because the profiled strips (i.e. showing a particular ‘stepped’ shape when looked at on end) are usually manufactured as solid strips rather than pieced together in layers and joining them on a diagonal is the only way of making the profiles fit. This also becomes important if the grain of the wood is to be a visual feature; joining on a diagonal is the only way to make this look good. In the model where, as here, it’s easier to build those profiles up in separate strips .. no, often it may be easier not to have to mitre, if the join lines aren’t too visible .. but here I’ve also chosen to put a fake grain on the Pvc, so mitre joins are better in this case.

Below, the easiest way to cut a 45degree angle is to use the grid of squares which can usually be found on the cutting mat, placing the end of the strip within a square and spanning from one corner to the other with the scalpel blade. Because the Pvc is soft the end could just be chopped by pressing the blade down, but I prefer just to make a guiding mark this way but then cut by drawing the blade across in the usual way. The other reason for this is that it’s better if possible to put a slant on the cut (shown by how I’m angling the blade in the photo) which will help in getting a tighter join later.

cutting a mitre

There is little alternative to using superglue when working with Palight foamed Pvc. One of the main advantages of using superglue is ‘instant gluing’, but anyone who’s used it knows that this can also become its main drawback .. the fact that one usually can’t slide a piece carefully into place and reposition it while it’s gluing. One has to develop little strategies to compensate. For example, rather than cut both top and bottom of the strip being glued in place below, I’ve found it better just to cut the mitred top and trim the bottom off (cutting from the other side) once glued in place.

gluing in place

The cutting and gluing of the pieces follows the form round, as below. Here I have cut a mitre on the second piece (also left for the moment as a longer strip) and am checking its fit before anything else. If this is fine, the position of the mitre at the other end can be marked in pencil while the strip is held in place. This piece has to be cut to size before gluing in place, whereas the third and final length can be left long when glued (as before) and trimmed afterwards.

moving round

Almost exactly the same is done for the window, remembering that the frame is cut off along the bottom edge of the window opening because the sill protrusion needs to go there (window frame wall mouldings don’t travel right round like a picture frame as a rule).

allowing for sill

Because Palight Pvc is relatively soft the surface takes grooving or embossing very well. Extra detail can be added to the frame this way if needed. Below, I’ve used an embossing tool (the tip is smooth and ball-shaped) to make an even groove without tearing up the surface. In fact by pressing lines carefully in the soft plastic in this way one can create the suggestion of a number of steps to the moulding without actually having to layer thinner strips on top. I’ve done it here (and it can be seen better in the last photos) because I wanted heavier shadows. The process is just the same used for the first layer.

embossing detail

Often in reality the whole window sill is a thick plank of wood which juts out a few centimeters into the interior space, the edge of which is rounded or at least softened. Here it is conveyed by a thin strip applied to the surface. Below, I am using the piece of 2mm thick Pvc left from cutting the door opening, rounding one edge first using the sanding block. When suitably smooth and even this can be cut off as a shaped strip, cut to length and glued in place.

rounding edge

The final part of this stage in the interior decoration was cutting another strip of 1mm Pvc for the skirting board. Before cutting the strip off the sheet I straightened and softened the top edge with the sanding block and embossed an extra line. Here I am using the sanding block upturned as a guide for gluing the skirting board strip right on the bottom edge of the wall. The sandpaper surface helps the block to grip in place for this.

using block as a gluing guide


The doorframe is always put in first and the skirting board inserted up against it. The doorframe is also usually thicker (at least at it’s outer edge) than the skirting board. The slope of the doorframe moulding i.e. just as a picture frame is thicker on the outer edge, is meant to perform the same function of drawing our focus in on the door. The skirting board has a few functions, mostly inherited from the past. It helps to define the wall, divide it and offer a visual transition from one plane or element to another, in the way that all other wall mouldings (doorway and window frames, cornice and dado area) are meant to do. But it’s principally there to hide the ragged join between whatever the floor is made of or covered with and the walls, especially floorboards where a gap is necessary to allow for wood expanding. It also used to function as a buffer, to prevent furniture from scuffing the walls.

completed wall decoration

Priming and painting the wall or adding wallpaper, adding door and window, making a shaped cornice, painting or staining the ‘woodwork’ etc. .. all these things will follow at some point ….

Making walls – Part 2

I couldn’t resist posting these results even though they’re hardly begun, because I’m pleased with them so far and wanted to at least make a brief record before I’m (as always) sidetracked into something else! I realized that, coincidentally, it has been exactly a year since I posted Making walls – Part 1 with the promise of a follow up .. so some attempt at continuity is long overdue!

decorated wall

I started working again on the two ‘wall’ examples here because I’m due to run my next ‘Scale Model-making’ course at Central Saint Martins and wanted to show some more evidence of what can be achieved with just Kapa-line foam. The wall structure above is pure invention or ‘architectural caprice’ with no particular style or period in mind, although it has turned out vaguely ‘ancient Greek/Egyptian’. The repeated decoration is very simply achieved .. using foam impression tools which in this case I’ve sliced from a set of plastic moulds used for cake decoration. I was only interested in parts of the shapes and in any case the complete shapes were too large to press into foam easily. The sliced details needed to be fixed (with hot-melt glue) onto wooden ‘push sticks’.

foam impression tools

I started the wall with the notion of giving it a heavy concave cornice but then decided to divide that into ‘teeth’ and to reflect those divisions in the rest of the wall.

building up wall decoration

These teeth (or ‘dentils’ as they’re referred to in classical architecture) are cut from the shaped strip using a try square to score an even line round and a long-bladed craft knife, as shown, to slice.

making dentils

The sides are neatened by sanding them against a right-angled sanding block.

sanding the sides

I use superglue to bond polyurethane foam parts together. Unlike other foams such as styrofoam the superglue doesn’t dissolve the surface, but it does soak in and start setting quickly (sometimes two attempts are needed). Superglues seem to vary and not all types work well with the Kapa-line polyurethane foam, but the best one I’ve found also happens to be the cheapest .. the ‘Extra Strong’ in tubes from Poundland!

wall in progress

Both the original cornice strip and the similar strip over the doorway were shaped using a sandpaper-covered dowel and the steps for doing this are worth noting.

setting up for sanding curved cornice strip

Above, I have sanded the edge of a piece of 10mm Kapa-line foamboard straight and perpendicular using the right-angled sanding tool shown earlier. I drew a pencil line to indicate the full width of the strip (i.e. not just the curved part). I then removed a strip of covering paper (by carefully slicing through the paper only and peeling away) exposing just the area to be sanded. Below, I’ve sliced away the corner along this strip, mainly so that the round sanding dowel will have something flat to start on.

chamfering edge

The sanding dowel is then dragged carefully against the edge, gradually creating and deepening the curve as it goes. With proper care a smooth, regular shape can be made in the foam fairly quickly. The edge of the paper covering left on acts as enough of a ‘stop’ for the sanding. Once this is done the strip can be sliced off to size (it’s usually better to extend the pencil line round to the back and cut halfway from both sides).

sanding a concave edge

Below is the other piece of wall I’ve been putting together at the same time, to represent the effect of old, sea-weathered wood which is achieved by dragging a wire brush firmly across the foam.

rough-hewn wall

For this I cut up some sample pieces I’d tried earlier, piecing together something less designed, more arbitrary looking. In this case I’ve basecoated the foam with a mix of raw umber acrylic and Paverpol, which gives it a much tougher/more durable surface without filling the detailed texture that much.

wire brush effect with foam

‘Making Walls’ will continue as a series ……..