When he was 16 years old, Harley Earl discovered he could make small cars out of clay by hand. Over a hundred years later, clay models are still an important part of the car design process and used extensively in nearly every OEM studio around the world.
These days digital horsepower is cheap. 3D modeling software is widespread, and additive manufacturing is available, so why do car manufacturers persist in using this seemingly archaic, expensive, and time-consuming method of making physical models? Wouldn’t it just be easier to get the pale CAD jockeys to do everything on a screen?
To answer the question, we’ll examine a model that Land Rover lent to Coventry University as part of a new exhibition celebrating fifty years of the automotive design degree, which after the opening evening may or may not have resulted in myself and ex-Jaguar designer chief Wayne Burgess quaffing beer in the local rock bar over the road after the boring speeches had finally finished. Made specifically for media and exhibition duties, the model perfectly shows the various stages involved in rendering a design in clay.
Before we start, something that I have mentioned before and is important to understand is that most (but not all) images purportedly showing the design process that are released by OEMs contain a certain amount of subterfuge. I’m not saying that manufacturers are talking bollocks, or the photos are manipulated, but what they show is only ever what they want you, or us in the media, to see. The design studio is the most highly restricted location on any car company’s campus – no one gets in who has no business being there. On those seldom occasions when outsiders (usually VIPs or journalists) are allowed in, they will have their phones confiscated or placed in a security bag that covers the lens. All this is a roundabout way of saying that the clay model we are about to take a close look at is a prop – it’s not an actual working model that was used as part of designing the Defender 130, one of the cars that I worked on and had responsibility for.
It’s All A Lie
The clay used is not really clay at all but more like industrial plasticine – the exact ingredients are proprietary secrets but it contains a small amount of sulfur giving it a characteristic stink. OEMs will get through hundreds of thousands of pounds of the stuff every year, recycling as much as possible. And it sticks to your clothes and gets everywhere, which is why I always hated making clay models as part of my degree course. Soft and malleable when warm, it hardens at room temperature but crucially remains workable and doesn’t dry out – so it can be milled by a five-axis machine and then finished by hand.
If the chief likes your sketches and decides your proposal is going to proceed to a full-size clay model, the first stage is an Alias modeler will turn your flashy Photoshop images into a digital model. This is the data that will then be the initial starting point for the clay. The model operations team then take that data and give it an internal structure using engineering software like CATIA. The clay model is built on something called an armature, an aluminum frame that allows for the wheelbase and track to be adjusted depending on the dimensions dictated platform. On top of this, the interior volume is filled out with wooden boxes, and then hard foam roughly outlining the shape of the car to be modeled, before finally a thin layer of clay is applied over the top. All of this needs to be designed and built before any exterior modeling can commence.
Why do it like this? You can’t make a model entirely out of clay: it would weigh eight tons and give the operations guys several hernias when they tried to move it. Also it wouldn’t survive the several years it takes for a design to go from ballpoint sketch to something you can walk into a showroom and buy.
Once the armature is built and the wood and foam core of the model added, several clay modelers will then apply clay by hand over several days in a process known as laying up. This takes place in a warm room to ensure the clay stays malleable by hand, so the modelers can ensure even coating, usually no more than about an inch thick. What finally emerges is a sort of car-shaped brown blob, like if you tried to sculpt a car out of a lump of poo. This will then be wheeled into the studio and onto a calibrated metal floor, known as a plate.
Cronenberg Could Do It
Depending on its size, a design studio will have a number of these plates in a row under horizontal strip lighting – at Land Rover, I think we had 24. Each plate has a five-axis milling machine attached to it, which uses the Alias data mentioned earlier to make several passes over the clay, starting out rough and getting progressively finer. The machine will run 24/7, and provided nothing goes wrong will take a couple of days to mill out your new design. You can see the various stages of milling in the photo below.
Once the milling machine has finished and the shavings swept up, a swarm of clay modelers will descend to finish the thing off by hand. These people are the real artists of the design studio. Using a variety of hand tools resembling something from Cronenberg movie, they scrape, sculpt and smooth the surface, adding details and finessing the surface so it starts to resemble sheet metal. If the model being made is an initial theme model (i.e. used to help decide what proposals to take forward) it may be lacking in detail – at this stage it’s more about the overall shape, volume and proportions. Things like lights and grills will be simple tape and paper mockups, because these graphical elements will be decided later as the design progresses.
So within a couple of weeks of having a successful sketch review and having your proposal picked, there is now a physical representation of your design, full size, standing in the studio. It weighs a couple of tons and costs hundreds of thousands of dollars to build, and the design studio of a major OEM will have several on the go at once. What is the benefit of this messy, labor-intensive, and smelly process? I mentioned earlier the idea of a working model. The beauty of working in clay is that it can be changed and altered as required as a design slowly moves through all the various milestones and gateways required to become an actual car. So a designer can look at a model that’s right in front of them, realize something is not working, and then get the modelers to change it for them. That change can then be scanned, and the digital model updated.
I Like Big Butts
To give an example of how this works, let’s go back to a car I worked on, the Defender 130. The problem was the 130 was never part of the Defender program from the beginning. When designing a car, right at the start of the process you try and think about all the future variants you might want to do, even if you don’t know if they will be released or not. The first generation Toyota GT86 was engineered for a convertible version that never happened. The Jeep Avenger has gone from EV only to having an ICE under the hood, like the new Fiat 500. It saves you a lot of headaches later on trying to do something that was never considered in the first place. With the Defender, the 90 and 110 versions were well on their way before some bright spark thought a three-row, seven or eight-seat version might be a good idea for the North American and MENA (Middle East and North Africa) markets.
Because the Defender was never meant to have a longer version, the shape of the car was completely wrong to be extended. It had too much boat tail – viewed from above the rear of the car narrows considerably because all the parts on the back are common between the 90 and 110. Adding another 290 mm or so behind the rear doors to create a 130 version meant a lot of work manipulating the rear three-quarter surfaces to make sure they didn’t look shit and could still use the existing tailgate and rear body parts from the shorter versions. We were able to get the clay modelers to sculpt that panel behind the rear wheels, have that scanned into data, and then get it milled to evaluate the surfaces.
In effect clay models are living, breathing representations of the car you are designing, morphing to reflect the changes made throughout the design process to make a design feasible. Remember that aluminum armature that underpins the whole thing? By loosening a few bolts that allow the track to be altered as different wheel styles are tried, tire profiles are worked out, making sure you have enough coverage to be legal everywhere your car is going to be sold. Because car design is a process that involves working with a constantly moving set of targets, having a full-size model that can be altered is crucial to seeing how engineering changes are going to affect the appearance of your car.
A further benefit of a clay model is it can be painted to look like a real car. Usually silver is the first color tried because it’s neutral and lets you examine the highlights – which is why the plates are all lit by lines of fluorescent tubes. The modelers will achieve a paint-like effect on the model using a vinyl wrap called Di-Noc, which is an industry-standard version of the stuff customizers use. As the final design progresses older clay models will be used by the Color, Materials and Finish (CMF) team to evaluate the paint. You can wheel them down to the paint booth, and a day or two later have something back in the studio that is the same color as what you’re hoping to sell.
So clay models don’t just give designers something to stand and point at for half an hour when they should be working – nothing can beat having a full-size representation of your design right in front of you. Digital models and screens are great, and useful tools but anything on a screen is just that; a representation filtered by the default camera settings in VRED. You can’t slap a tape on them or make them look like a real car that visitors to the design studio will be wowed by. Part of design is salesmanship, and nothing makes those visitors to the studio understand what you’re trying to do than a big full-size clay model they can see and appreciate.
The Pontiac Aztek Was Not A Design Tragedy, It Was A Corporate Tragedy.
The Land Rover Defender Outbound Has One Infuriatingly Stupid Design Detail
Here’s Why The Supercharged V8 Land Rover Defender 110 SE P500 Is The Goldilocks Land Rover
I have a question on design perspective. Or more literally, point of view.
A great number of photos of cars, both in advertising and in car reviews, are taken with the camera between hip level and knee level, and occasionally even lower. I understand that this perspective can appear “dramatic,” “commanding,” active,” or various other wonderful marketing terms. And so I understand why the adverts use these angles.
What I have less understanding of is why reviewers use these angles. Is their goal also to make it look exciting? I can state that my personal goal in seeing pics of a car that I’m interested in and can’t see in person, whether because of rarity or because it’s not on the market yet, is to know what it looks like from a few meters away with the camera at eye level, from 45 degrees off the front and rear quarters. I wish reviewers would routinely provide these normal, natural, human angles.
So my question with regard to this article and your experience, Adrian: when designers are working with the clay model in its sulphurous flesh, are they ever or often crouching down or taking low, low photos to evaluate from the perspective of a cat or small to medium sized dog? To what extent are designers designing for the advertisement perspective, versus for a realistic street view?
One of the benefits of working with physical models is that you view them with a human eye – ie exactly how a car will be viewed in real life. Designers do look at things from close up, or from odd angles but that’s to make they haven’t inadvertently made a mistake. Designs are not evaluated from a worms eye view, or dogs eye, but in some cases you might evaluate from a different angle. For example when I worked on the Defender 130 we had a lot of work to do getting things like the roof rails, rear spoiler and infill panels right because of all the carryover parts from the shorter versions. So we spent a lot of time on the mezzanine looking down onto the roof of the model.
Studio photographers are there to document the design process and capture images for later review. They will normally take the standard front, side, rear, front three quarter and rear three quarter shots. Press images are all about drama so will aim to show the car as dynamically as possible, so you might get some funky views then. But really to properly understand a car you need to see it in the metal (or clay).
So then, pretty much as I thought. Designed considering normal viewing perspectives, not designed to advertising perspectives, and the marketing photographers can do as they will after it’s done.
Before I posted the question, I checked the manufacturer’s site for the most normal car I could think of, the Toyota Corolla. It is not entirely crazy angles, though there are plenty of them including in videos; at the same time many of the normal views are still unrealistically low, e.g. the side views are shot from door handle level.
Always worth checking the media images.
Great article! As an engineer (not automotive) I have spent countless hours in CAD software and you never, never get a true sense of how something looks in meat-space until you have a prototype or 3D printed model or something physical. It’s really irreplaceable.
Very, very interesting, as always, Adrian. Makes me appreciate things like the Jag E-type even more now, thinking about how that was designed out on the lawn during tea or something. But did Lyons do the jam then cream or the cream then jam on the scones?
Now I’m thinking of how Saab did the ur-Saab probably without a clay model, just kinda winged it.
> like if you tried to sculpt a car out of a lump of poo
Fittingly, British carmakers do this in reverse at the factory.
Great write up. I’ve had the same experience with 3d printing. As an engineer, you can get lost in the CAD model after a short period of time. I’ve gotten in the habit of 3d printing regularly throughout the process. Obviously doesn’t work for a full size car, but for small things, or scaled down things, it’s simply amazing.
As a home hobbiest and avid DIY’er, I will regularly make one or two changes a day, slowly iterating a design over time until I’m done, while printing the changes each day. Another word for in in R&D is prototyping.
One of the 85 projects I have on the go (but not actually progressing) is getting smaller scale 3D models of my graduation projects made. I still have the hot rod and hope to have the Jeep model back from the RCA soon.
I’ve always wondered – did they ever use actual clay, or was it the industrial stuff from the very beginning of its use?
Like I can imagine the studios that produced various Morgan designs boasting of models represented in the finest English clay or somesuch…
Recycled Gumbys and Pokeys.
I admit I’m not 100% sure how it was done in the very early days. Earl wasn’t really fond of modelling as a development and experimental tool, he used it more to confirm things and get the engineering done, so I’d have to check.
How did they shape it before they had 5-axis CNC mills?
Before that, how did they lay out the design before 3D CAD? A: full-size drawings on plastic film (matt, a little like thick tracing paper) using various thicknesses of flexible black draughting tape: “tape drawings”. Usually a side view plus the main sections.
The film had a grid on it (100mm squares) with one each of the vertical and horizontal axes running through whatever the 0,0 axis of the particular maker was, from which the modellers could measure off data points. Centreline, belt line, feature lines, glass openings, sections, etc. would be picked up along the tape lines.
Those points would be picked out on the lump of poo model mentioned by Adrian, using manually operated 3-axis measuring machines running on rails on each side of the plates.
The points would be highlighted with Tipp-Ex, then the surrounding surfaces scraped away with the Cronenburg tools down to the level of the points, leaving just a tiny white Tipp-Ex dot. That’s it in a nutshell. After that a lot of chin-clutching, redoing, checking hard points, refining surfaces and lines, etc. etc. & so on.
I am so old that when I first joined an OEM as a design intern, they were still doing body drawings for tooling by hand, measured off the final clay model. In those days, the final model was an important engineering tool once the design was signed off.
So old the car’s horsepower came from oats.
It was a long time ago, but not so long ago as you might think. The changeover from analogue to digital starting at the end of the 20th Century was mind-blowingly fast and is still accelerating.
Harley Earl was famous for insisting the modellers had orthographic drawings before they started work.
Adrian, that’s before my time by about the same amount as tape drawings and manual point-taking are before now. Also before the time of Mylar and graphic tape, so I guess the designers did full-size drawings on huge sheets of paper using pencils and sweep and ships’ curves?
Before my time a well, but yes full size renders done with airbrushes were a thing. If you go back and read the Damn Good Design about the ‘77 Chevys there are some studio shots that show full size renders.
When I was in design school, one of my manufacturing professors brought in a set of blueprints from a C4 Corvette. One sheet was about 10 feet long and had nothing but full-size cross sections of the rear hatch glass, sectioned every few inches. These would have been used by the clay modelers as they hand sculpted (back in the late 70’s/early 80s) and by the tool makers.
fascinating explanation. thanks for sharing. I had thought that clay had been replaced by tech
Supplemented, certainly, but never replaced
One of my greatest regrets (other than the obvious) about not pursuing a career in auto design is not working with clay. I adored doing ceramic work in HS and during my university time pursuing an art degree, especially putting my colored pencil drawings into 3D reality. It’s such a pleasant medium to work with, if you have good product and know how to take care of it.
Peruse the machinery auction sites looking for used Tarus claymills.
Try to think where I would put it.
Sigh.
Clay mill tool life must be the envy of machinists everywhere.
Wow! Talk about feat of clay. This is a great write up about a fascinating process. Except that manufacturers don’t want us to see it, this would also make for a great time lapse documentary. Thanks for the insight.
I suspected that would be the answer, along with the idea of being able to see and touch an actual object (even before applying paint, etc.). When I’m making brackets and such, I’ll generally make a sketch and then create the initial prototype in thin aluminum stock which is very easy to shape, even by hand. Then, when it’s as correct as I’m likely to get it, the final version is recreated in steel.
Fun fact: hard foam like that can be shaped very rapidly using an extremely coarse file-like object called a rasp.
If that hard foam is the same as is common in thermoforming prototype tooling, it’s nasty stuff. Sand-like detritus gets everywhere, wearing cover plate slides in the CNC, and a high rate of wear on end mills. I received a box of the adhesive, read the MSDS, and told the head of market development it would be foolish to open it.( in case of accidental spill, evacuate building) This was a small R&D lab with 36″Z, 40″Y, 80″X CNC, thermoformer, curing oven, end mill, band saw, closed off computer office, all across the back of the U.S. headquarters for an international EPP supplier. I convinced them that we should be using our free to us EPP instead. Worked out great.
That sounds a little harsh for something used in auto modeling – dang. 😮
According to the pics, the foam appears to be supplied in blocks which are then assembled (with some sort of adhesive) to create the larger shapes. How is the thermoforming version shipped?
Fun fact: I googled for “thermoforming prototype tooling” and your post came up. 😀
That’s odd. I remember the smell, feel, color and weight of the stuff, but not the name. It looked exactly like the pictures here, including the discolored lines where the penetrating primer/epoxy joins blocks. You need an air supply suit and sealed room to apply safely. It turns to sand like dust in the CNC, and you can program aggressive machining, but a surprising amount of wear. The blocks shipped regular commercial. the epoxy had lots of hazardous material labeling.
When I first started working with the hard foam, we ordered blocks big enough that no joining needed. And now that I am recalling from 28 years ago, I think the high hazard stuff was to put a hard, heat resistant finish on it. Though a lot of research, I found a non-toxic, water-based, 3 part epoxy that was developed for coating antenna towers to sheet off ice, called the company and got the owner, told him what temperature resistance I’d need, he sent me some samples to try out, got a great result with his standard formula that gave it a soapstone like finish, and you could fine sand it, and re-apply if needed.
Interesting – thanks for the info!
Of all the bizarre career positions I’ve had, that was the one I enjoyed most. Thought I would be able to retire with them after getting some patents for the company, but when the head of engineering retired, there was a political power struggle, and the new previous sales head somehow got the position, got me fired after some disagreement, and he was booted 6 months later.
This is the stuff I made my two big models from:
https://www.neillsmaterials.co.uk/product/sikablock-m330/
Whew – that appears to be much safer to use/work than the stuff Hoonicus was describing. Thank you for the link!
I seem to remember the adhesive was definitely of the don’t sniff in an enclosed space variety, but didn’t need a hazmat suit.
Occasionally a rough speed foam model will be quickly milled out full size, but I saw this happen only a couple of times.