The polysytrene used was high density white closed cell - about 30 kg/m3 which the supplier, Styrobeck, hotwired into 25mm thick blanks. It took a hard thumb push make to dent - something guaranteed to ban the offender from the workshop.
Safety note: polystyrene should be treated with fire retardant, an important consideration when large volumes are stored/used in a home workshop. And too much exposure to styrene fumes has the hallucinogenic properties of lysergic acid and the emetic properties of cheap red wine. Just to spoil the fun the latter comes first, be warned...
After 200 hours the first set of slices stood ready for lamination.
Particle board uprights and a dual threaded rod armature were essential to keep the laminate sandwich square, correctly dimensioned and under pressure until the PVA wood glue set.
More slices sit to the left of the photo, ready to be laminated in place, but first housework - endless sweeping. Can you spot where the software made a mistake with the tool path?
String lines helped establish dimensions.
Note that the dreaded yellow handled Stanley Surform now enters the picture. A device that in 10 seconds can wreck the work of 10 hours.
Advantages
Direct translation from CAD.
Detailed curves, both convex and concave, were shaped with ease.
Cheap: especially when compared with polyurethane - admittedly a better product to carve and sand; or multiple sheets of ply/particle board - which though accurate still needs a foam or chicken wire and plaster surface.
The car grows before your eyes.
Tea drinking while the cnc cutter does its work.
Get low enough and it's even possible to imagine you're sitting in it!
Disadvantages
The crude nature of software meant that it occasionally thought a straight cut was the shortest path between two points. Ouch. Also it sometimes plunged straight down when it should have navigated to achieve an undercut. Ouch again.
The minimum concave curve was 26mm diameter while the minimum undercut was 10mm. Generally the job could be arranged so that this wasn't a problem. But all hell would break loose if a mistake lead to the spindle of the cutter being dragged through the job.
The wood rasp cutting head was no laser and left a layer of 'fur' on the finished job. Moreover, after one or two hours polystyrene would melt into the cutter teeth, which mean that the cutter then forced its way through the job and occasionally ripped out chunks of polystyrene.
Steps and stairs remained that had to be Surformed or filled at a later stage. With no hard templates as a guide it was too easy to Surform a 'flat' into the buck that had to be filled later.
These problems could have been fixed by building a better cnc machine but this would have cost time. In fact, some of the rough and gouged trial pieces were still used in the buck because close enough was good enough.
Several hours careful work with a Surform and some infill polystyrene and expanded foam produced a buck that could be brought to a more regular and hard surface.
The book, Fiberglass& Composite Materials by Forbes Aird was a useful reference at this point even if Forbes spells Fibreglass incorrectly. The Fibreglast site is also worth a look, especially on the topic of mouldless construction.
The next stage was damn unpleasant – but after two years the nightmares have stopped and I can, at last, confront the memories of building the buck (or plug).
All I had to do was make a 1:1 movie prop in my workshop… easy huh. Forbes Aird explains the technique in his book “Fiberglass & Composite Materials”. The idea was to surface the polystyrene with a hard shell of body filler.
Bog loves to eat polystyrene. Several test strips showed that about three coats of water based house paint gave good protection. At last a use for all those half tins of house paint with their evocative interior decorator names, Half Tea, Colonial Cream, Sandfly Bay Gray…
Picture 1 shows that I chose this point to cut out the engine cover so that I could construct the lip that goes under the cover. Progress was good. The polystyrene was a pretty Brunswick green. Things seemed so pleasant. Even the children could be bribed to help.
The mathematically inclined may already have seen the horror approaching. The buck had a surface area of roughly 10 sqm. If I wanted just 2 mm of filler for the surface I’d need 19 litres of filler not counting what I sanded off. Cripes I thought, at about $30 per tin that could cost a lot of money.
Surfacing also meant inserting hard edges into the polystyrene at key points. The dashboard was bull-nose routed from a slab of chipboard, inner wheel arches were shaped from alloy. This was intense stuff often late into the night.
A moment’s inattention put a box cutter through my leg. I didn’t really notice until my gumboot got warm and wet.
I coloured the first layer of bog with red oxide and combed it on with a grout spreader. When this was hard I added another layer of plain bog. That way if I sanded back too far strands of colour would appear and get wider. Working in anything less than a strong northerly caused the house to reek of polyester and the family to fall about complaining and puking. Children now had to be dragooned to help.
Lots of bog was required where the cutter had gone feral including a hollow 15mm deep where the CAD had given the buck a waist (you just can’t see this sort of thing on a computer screen). All in all it took. 13X4.5 litre tins of bog – that’s 52 litres of bog. You do the costing, as I can’t bear to think about it.
I made up a 1 metre ‘long board’ so that the car’s sides would be dead straight. And a shorter 300mm sanding board for convex curves. The key to convex curves is to sand in a diagonal using an X pattern – back and forth sanding ensures that you’ll cut a 'flat' and I’ve cut a few. Concave curves required a variety of curved and rubber sanding blocks.
Sand, sand, sand, fill, sand and sand. There are no photos of this horror stage.
Hours of labour lead to a rotator cuff injury.
The compound curves were tricky. Long thin plastic strips helped to gauge whether the curves were fair – did the strips sit evenly on the curve of the body. A cheap rotating laser level helped to mark bumps, levels and centrelines. Lots of bobbing up and down helped my eye follow the ‘horizon’ of the curve although this also required the buck to be one colour.
Then wet sand, sand, trace colour, sand, feel the curve by hand, sand and sand again. Months went by dressed in overalls and a filter mask.
Steroid injections were now required for nasty case of tennis elbow.
The buck didn’t have to be perfect. After all, one stage or mouldless fiberglass construction meant that the outside of the buck would form the inside of the fiberglass shell. But the gross shapes had to be right, especially the head fairing and headlight rebates. Frankly, I believe you could give one side square wheel arches and the other side round and no one would notice, but everyone would notice if the headlamp rebates were asymmetric (or cross eyed, Mr. Morgan).
This was also the time to split the buck into the single piece front and the rear clam shell. If I cut them neatly, I reasoned, then I would have perfect shut lines in the fiberglass parts (you will gather by nature I’m an optimist).
This lead to one of the psychological highlights of the build. I had wondered how Porsche had drawn the curve at the front of the clamshell – which cuts into the leading edge of the rear wheel arch in Pic 4 above. French curves? Flexible ruler? Detailed measuring? In desperation I put my elbow on the wheel arch and pivoted my forearm like a compass - freehand. The spirit of Ferdinand Porsche flowed through me.
The result was perfect. This must be how Ferdinand did it fifty years ago.