How are the prototypes for plastic kits made?

[Ben]

Watch this video series:

Cool! Very informative!

[Art Anderson]

One could almost write a book, in multiple volumes about how a model car kit gets created, but I'll try to be as brief as possible.

The process starts (once the "Go" decision has been made for a particular subject. Now, here's where the first crunch can come! As late as the 1980's, automakers were still using what are called "loft drawings", a type of drawing pioneered in ship building at least 250 years ago. This type of drawing is of course, on flat paper, whether it's created by CAD, or by human hands at a drawing board, using drafting techniques which show the curvature and depth of the various 3-D surfaces of say, the body shell. So, back "in the day", model company pattern makers had to know how to read those drawings, combining that with whatever photographs could be gotten or taken of the actual subject. In the heyday of promotional model cars (where our hobby and it's industry really took off!), that information came from the various automotive styling departments, well in advance of the introduction of the new car at dealerships.

But, many of the cars we really groove to are old enough (and in a lot of cases, the companies that made them are either no longer around, or if they are, archived drawings and such can be very incomplete, even non-existent--or at least their existence can be unknown to anyone even within the particular automaker). So, this is where the research really begins. It involves finding such images of the actual car as may have been published (established model companies are known to have considerable libraries of car magazines, books (from the technical to the "coffee table" kind), simply on the strength that someday, their product development people may well need them. Next, whenever possible, an example or examples of the real car must be found, that can be photographed. Now, the pictures taken for scale model development are not limited to the kind of beauty shots prized for publications, but rather, are taken to show as much as possible the actual car from every conceivable angle, even concentrating on details. In addition to high quality camera's, the tools of this process include such mundane items as tape measures (both steel tape and soft PVC coated cloth tape (like any seamstress would use in measuring fabric) where a measurement has to be taken up against a nice paint job. A general tradition here is to take a permanent marker, and black out every other inch, so that once the pic is taken, measurements can be read easily (remember, in 1/25 scale, one inch equals .040", or very slightly less than 1mm). The photography process will include shots straight from each side, one from the front, one from the rear, and a series of 3/4 shots all around the car. Whenever possible, development people like to have a tall stepladder, or other means of getting well above the subject, to capture views from above, which while not often seen of the real car, tell the designers and pattern makers what the car looks like from a modeler's eye. Add to this, perhaps several hundred detail shots, showing shapes and dimensions of such things as windows, vent wings, shapes of taillights and headlight bezels, cross section related shapes (when I was developing Johnny Lightning diecasts, a standard part of the photoshoot included opening the doors, and capturing the shape of the door from the rear edge, as that showed the mockup makers the contour of the body sides, in the same manner that "station section" cross section drawings) would show. Every detail shot generally will have one of those various measuring sticks or tapes laid on, or held next to, the actual body surface. This also helps in getting the correct location (vertical and horizontally) of things like chrome trim, door handles, badges and scripts. In addition, a lot of shots get taken at severely oblique angles to the subject, as that can show further the body contours, and even the contours and depth thereof of things like wheel rims, hubcaps and wheel covers. Lastly, whenever possible, they like to be able to get the actual car up on a lift, to get the underside--this can take a lot of frames, in order to get the whole picture!). Along the way here, lots of notes get written down, hopefully to not miss anything that might prove very important later on in the process.

It's the same with the interior, particularly if it's a long, out of production make, model or body style. Pics get taken through the doors, the dash from as many angles as possible, even from over the back of the front seat while sitting in the rear. Pics of engine bays get taken, to show as much of things like inner fender wells, firewalls, radiators and their core supports, all of that. This information, when combined with service manual pictures and drawings can go a very long way in getting the model right. Engines and transmissions of course can be worked up from service manuals, even junkyard or swap meet pictures.

From all this information, the pattern or mockup making can begin. Prior to the advent of CAD/CAM, every part of a model kit had to be mocked up, generally in a soft, smooth-grained wood such as basswood or clear white pine. This process alone required sculpting skills rivaling any artistic sculptor of statues and the like--because in order for the model to look right, the patterns or mockups have to be right from the get-go. Time was, when tooling mockups, thus created by hand, from wood, were made to a much larger scale than the production model kit: Generally speaking, companies doing 1:25 scale kits started with patterns done to 1:10 scale, while 1:24 scale kits were mocked up in 1:12 scale, both scales being readily reduced by simple math. These mockups, once approved, were then cast in resin, from which resin female castings were made, to be inserted in large scale wooden patterns from which the dies were cut. The dies used to be cut using rotary mills, set up with 3-axis pantographs that could move the cutters front to back, left to right, and up & down. They were manipulated by hand, the toolmaker moving a stylus over the large scale pattern of the tooling. That was a laborious task, took several hundred man-hours to finish. Once the steel dies were ready, they were placed in a mold machine, and plastic shot into them, for the first test shots. From those test shots, things such as material thickness (I've seen early test shots that had thin spots to the point of their being actual holes in the body shell!), mold alignment and such. Once all that was checked out, corrected wherever needed, then the fine surface details were cut into the steel--things like scripts, badges, door handles and the like generally requiring a manufacturing jeweler's touch. Once that all was checked out, test shot, if the dies were ready, they were given a final polish to a mirror finish so that the finished plastic parts could come out as we modelers like to see them.

Of course, CAD/CAM, and laser technology have replaced most all of the labor-intensive manual operations in the process--nowadays, many model kit tooling mockups are made with computer aided technology, and the rotary milling cutter has been replaced by electro-discharge machining or EDM) which uses precise electric arcs to do the cutting and shaping of the tooling. And, of course, laser scanning and CAD have replaced the old timer's drawing boards, with their accompanying T-square, triangles, dividers, compasses and drafting pencils and pens. But still, GIGO (Garbage In, Garbage Out--which probably every student whose ever taken a programming course has heard sometime) means that just because the computer says, it may not be actually correct. This is as true of CAD files as it is of laser scans--I've seen both that were way, way off, particularly given that the people overseas who do this work simply do not have very much (if any at all!) personal experience with the cars they are expected to create model kits of. BTW, rapid prototyping, as in 3D printing seems not to be much used in the making of modern tooling mockups, but rather CAD is used, starting with solid blocks of material (at least, the tooling mockups I have seen show no evidence of any of the layering that can be seen to at least some degree in 3D printed objects). But, for sure, it is quite possible for tooling mockups that are made from computer aided machining to be incorrect at least in some areas--that is one of the reasons one sees anomalies in the tooling mockups Dave Metzner so graciously has put up on this forum. Those things take time to get corrected, to get them right. Of course, language and cultural barriers can play a pretty big part in all of this process as well. What almost nobody outside of the process knows is, that when Dave got what were essentially correct body test shots of the '55 Chrysler, there was still an incorrect contour on each side of the roof between the B-pillar and the back window. It was a simple job for me to take one of those test shots, build up the offending area with catalyzed putty, and with file and sandpaper, get the contours corrected (did that in about 3 hours!). And, I think the finished model kit looks pretty right in those two areas.

But, the bottom line is: It's never been a "snap your fingers" simple sort of thing to create a model car kit--whether it gets done the old fashioned way I've described, or by modern 21st Century technology, it is still a very complicated and complex process, which can take the occasional twist, turn, and perhaps the wrong "fork in the road". In short, it takes not only a lot of high skills, but also dedication and a passion for seeing the project come to fruition in (hopefully!) a timely fashion, and within budget. When all that comes together, the resulting kit generally winds up being a winner, pleasing the vast majority of builders.

Sorry for the very long post, but I feel that this is a story that needed to be told.

Art

[philo426]

Can you imagine the tooling required for a detailed WW2 aircraft!That would be even more complicated! Especially if it has folding wings!

Edited August 5, 2013 by philo426

[Art Anderson]

Truthfully, while the assembly of a WW-II naval aircraft kit having movable folding wings is rather complex, the tooling for it really isn't. Parts such as those are molded in fairly simple molds--a right side and a left side.

On the other hand, a one-piece model car body requires rather complex tooling--as almost always, 6 separate dies have to come together precisely to mold one. These tool sections are all individually movable--the mold sections which create the exterior of each side, each end of the body shell must move in and out with each cycle of the tooling, along with the mold having to close and open with each molding cycle. Just imagine not only choreographing that "dance" while at the same time achieving the precise mating of all the mold sections (called slides, or sometimes, sliding cores) each and every time.

Art

[philo426]

I hear you Art!Did you ever notice how the body shells have uniform thickness?Not an easy task to accomplish.

[wisdonm]

Thank you for that explanation Art. It's the most thorough one that I have read yet. The pantograph explains a lot of things.

[clovis]

Art,

Let's say a model company wanted to make a brand new model.

Do those companies try to find original blueprints?

Do they ever contact the manufacturers, like GM, for detailed information?

It seems that a set of blueprints, or the equivalent in CAD, would make designing models so much easier.

Also, I always envisioned that model companies used giant, but lightweight calipers to measure body width at certain points on 1:1 examples. Does this ever happen?

Do model companies always use 1:1 examples for measurements, or do they SWAG it sometimes?

[Art Anderson]

Jeff,

Original factory blueprints have rather limited use in working up a model, if for no other reason than their sheer size, and that most are of mechanical or structural components--such drawings as would have depicted how say, a body panel was to be shaped, probably dealt far more with the stamping dies of which there can be many--some body panels back in the 50's and 60's required at least a couple of dozen stamping dies just to press out a quarter panel.

From perhaps the late 1920's until well into the early 1990's, car bodies were styled first by artist's renderings done on paper by stylists, then actually sculped in a type of modeling clay in actual size (the late Gordon Buehrig, who styled numerous bodies for the legendary Duesenberg Model J, and later the futuristic (for 1935) Cord 810, actually worked in 1/4 scale clay models, but full size clays quickly became the norm. Buehrig also is credited with inventing what was called the "styling bridge", a wooden frame that actually spanned the clay model, with holes drilled in it fairly close together through which round dowels (probably metal) could be inserted to contact the clay surfaces, with the distances between the styling bridge uprights and cross beam that spanned the top of the clay body could be noted in order to translate that shape into information that could be then translated to creating stamping dies. That's pretty much the same principle in which a modern laser scanner works, the laser beam being measured to get the same information. So, there really weren't "blueprints" that could give a model kit developer much in the way of really workable data.

In the days before laser-derived information and CAD/CAM, doing model kit parts in the tooling mockup stage for a model kit simply required the services of highly skilled artisans--sculptors in every sense of the word--no less artists than a Michaelangelo (who created some of the most realistic sculptures of the human form in history. In short, the ability to translate what the eye could see, to carved 3-dimensional shapes was vitally important. Of course, there were full size plan form drawings created by automobile stylists and their draftsmen, but those served to 1) show the lines of the car in side, front, rear, and top views, and to provide documentation for patent applications and copyrighting.

As for working with the actual manufacturers for detailed information--absolutely, whenever possible that is done. But that is dependent on a couple of things--is there any original factory information still available, or is the company who made the actual car even still in business. Complicating this (and the Moebius Hudsons come to mind here) is the fact that many smaller, independent carmakers never had their own body manufacturing operations, particularly as automobiles and their bodies became more complex with the advent of all-steel construction and streamlined, fully enveloping body styles. In the 1930's, there were two very large body companies operating within the auto industry, Murray, and Briggs, both of whom supplied not only smaller companies such as Nash, Hudson, Packard (who was never a huge producer of cars BTW), and so on. Most independent makers may have had a stylist or two on staff, but once a style was approved, the real work was farmed out to them. In addition, even Chrysler Corporation didn't have their own body manufacturing division until the corporation bought out Briggs Body Company in 1954. Couple that with all the changes and shuffles within the auto industry over the past more than a century, with companies being formed, closed down, merged into other companies or simply bought out, AND the necessary concentration on not just what they were doing on any particular day and year, but also what they were going to do in the future, not much emphasis was placed on saving history, at least not to the extent of saving every scrap of drawings.

As for working from sets of old blueprints--most of those would be cumbersome as all get out, to put it mildly--given their sheer size (I saw Lesney AMT designers working on the the '67 Camaro that they released in 1980 as an all new tool--and they had a 1/4 scale side view "loft drawing" tacked up on a very large easel, that was clearly a GM Design drawing--but the rest of the work appeared to be in progress working from extensive photographs. Of course, with CAD, and laser scans of the actual car, this process becomes a lot easier--but so much depends on not only the person at the computer, but also the software and the quality of the scans. Even with that, anomalies do come in--the finished tooling mockups simply have to be studied, and more often than not, corrections get noted manually, and they have to be made, by whatever processes or technologies that get the job done.

Huge calipers???? Seriously, I cannot imagine anyone lending their collector car to be measured and documented for scale modeling allowing anyone to put such tools on their pride and joy--can you? (Frankly, when I was creating my own masters for resin casting--I'd jump at the chance to measure and photograph a rusty old junker--simply because I could do the work without fear of scratching the thing--and worn paint, rusty surfaces show the shapes and contours far better than a 100-point shiny restoration!) Actually, if the actual basic dimensions of the real car are known, it's not always necessary to get every single dimension exactly perfectly--in 1/25 scale, for example, a quarter inch is a mere ten-thousandth, or about the thickness of a cheap business card or the Ace of Diamonds in a deck of cards. But, I have stretched a measuring tape through the interior of a 2dr or 4dr sedan a time or two, from the surface of one side of the body to the other (with a helper, of course!) to determine the widest point of the body--works quite well when the sides of say, a 60's car are virtually straight and parallel. And of course, dozens of photographs taken from all manner of angles, concentrating in this or that shape, this or that portion of a body--just about every bit of photographic information you can imagine there, are extremely valuable to the person creating the actual model kit body in mockup form. Be the work done the old fashioned way, or by modern technology, all of that can be vital.

Also, if one is doing a model of almost any American car from the 50's through the beginning of modern "jelly bean" styling, simply knowing how the actual car was built, what body panels were common across a line of cars (for example, '59 Chevies all used the same front clip, variations of the same rear quarter panels, a couple of different trunk lids, and of course, different roof lines, but the same basic door skins were used on all 2dr bodies from Chevrolet that yea4, be they 2dr sedans, 2dr station wagons, convertibles, hardtops etc.--and two different windshields (framing and glass) (2 and 4dr hardtops used a lower, more sharply raked windshield than say, a sedan, station wagon or El Camino, due to the lower roofline of the sportier body styles)--all this knowledge has great value.

But, the bottom line is: if one is to do a scale model of a car--almost always much of the work is done from actual examples whenever possible--that's still the gold standard, the best way to go about it. For that reason, it seems that in this country at least--model car kit development gets done by people living in fairly large urban areas, and certainly where museums and collector cars are readily available. It's much, much easier to find examples of the real thing existing within a couple hour's drive, if you live in the Great Lakes states, the Northeast, or say, in the Los Angeles megalopolis than to try doing it while living in East Undershirt or West Overshoe miles from nowhere.

Art

[Art Anderson]

Here's the reason not many full drawings exist of a fully completed automobile that show all the intricacies of the body shapes! This is a recreation of the original styling bridge that was invented by Gordon Buehrig, when he was styling what became the 1936-37 Cord 810/812. If you look closely, you can see how dowels, graduated in fractions of an inch, could be inserted into holes in the wooden frame, and pushed in to just touching the surface of the model--from which the distance of each point on the Cord body could be noted, along with its position measured from front or rear, in order to give the information to the pattern-makers responsible for sculpting the stamping dies for producing the body panels. In short, the information was far more numerical, and a lot more useful, than any blueprint or drawing could have shown.

This technology served the auto industry all the way out to the inception of computerized technology--a laser scanner does pretty much exactly the very same thing, much faster, and hopefully, much more accurately:

Art

[Ace-Garageguy]

...from which the distance of each point on the Cord body could be noted, along with its position measured from front or rear, in order to give the information to the pattern-makers responsible for sculpting the stamping dies for producing the body panels. In short, the information was far more numerical, and a lot more useful, than any blueprint or drawing could have shown.

Though I agree with virtually everything Art has said here (very clear, well-written and informative, Mr. Anderson) I'd like to point out that in most cases, the dimensional data derived from the "design bridge" measurements of a clay were turned into what's referred to as "station drawings", which art mentions earlier. These are simply slices at specific intervals of what the car would look like if literally sliced into at that point, and rotated 90deg. towards the viewer. Naturally the stations have to be closer together on areas with more fine detail, or that are changing rapidly. The station drawings were the actual method used to transmit data to pattern makers in most cases.

The processes involved are very similar across all industries manufacturing familiar articles from phones to bottles to cars.

This full-scale model is being developed from a 1/10 scale original, with 10" "stations" clearly visible on the large drawing in the background. The stations were used to create full-scale ribs, visible toward the rear of the vehicle (using a tool very similar to the bridge shown in Art's photo) and the areas between filled with foam. The final surface is developed with fiberglass and bondo, and production tooling pulled directly from that.

The original model in 1/10 scale is my avatar, which I designed in 1984-5

[johnbuzzed]

With all the info which we have been provided in this thread, it amazes me how and why manufacturers can manage to mess up, a la the new Revell Mustang or the front axle on the Rat Roaster . Lost in translation, perhaps?

[Ace-Garageguy]

With all the info which we have been provided in this thread, it amazes me how and why manufacturers can manage to mess up, a la the new Revell Mustang or the front axle on the Rat Roaster . Lost in translation, perhaps?

it really all comes down to every member of the team caring enough to do the job he's paid to do, every day.

Guys being paid to measure need to measure accurately.

Guys paid to input data into CAD need to input accurately.

Team managers getting paid to catch errors need to see that mistakes are corrected BEFORE the commitment is made to cut tooling.

We have the means to electronically transmit ANY project or tooling data to and from China instantly, so that's no excuse so long as EVERYONE in the chain does the job he's paid for.

It's pretty simple.

[Art Anderson]

Though I agree with virtually everything Art has said here (very clear, well-written and informative, Mr. Anderson) I'd like to point out that in most cases, the dimensional data derived from the "design bridge" measurements of a clay were turned into what's referred to as "station drawings", which art mentions earlier. These are simply slices at specific intervals of what the car would look like if literally sliced into at that point, and rotated 90deg. towards the viewer. Naturally the stations have to be closer together on areas with more fine detail, or that are changing rapidly. The station drawings were the actual method used to transmit data to pattern makers in most cases.

The processes involved are very similar across all industries manufacturing familiar articles from phones to bottles to cars.

This full-scale model is being developed from a 1/10 scale original, with 10" "stations" clearly visible on the large drawing in the background. The stations were used to create full-scale ribs, visible toward the rear of the vehicle (using a tool very similar to the bridge shown in Art's photo) and the areas between filled with foam. The final surface is developed with fiberglass and bondo, and production tooling pulled directly from that.

The original model in 1/10 scale is my avatar, which I designed in 1984-5

Actually, I believe the concept began with ship-building, probably sometime in the 1700's, when sailing ships began to be designed with building multiple ships from the same design. Certainly, ships today are designed in that fashion--if one looks at any of the more technical histories of say, WW-II US Navy ships, generally books on a particular ship will show those station drawing cross-sections. Given the sheer size of any sea-going vessel, to have tried working with full size drawings of the actual ship would have been so cumbersome as to have been very impractical. Also, by the advent of steel ships, hull design became very important, so scale model shapes were tested in water filled basins, with the model hull being drawn across the water surface--the purpose being to create stable, fast and yet maneuverable shapes.

Art

[1930fordpickup]

The Pattern Makers need all the prints with good dimensions to produce good parts . Yes I have been there and done that . We would make our wooden models and than the customer would eye it up for the big OK . If it passed we would make the blow molds . Not the same as an Injection mold but to get to the point of the mold it is close to the same steps.

[Bernard Kron]

Thanx so much to everyone who has contributed so far to my initial question! There’s so much rich information here, both explicit and implicit. Casey, the FineMolds videos are fascinating, a great insight into mold making as it is done today. And the photos of the wooden bucks that Bill Wowk and Casey provided are very impressive and bring to life Art Anderson’s comment: “In the days before laser-derived information and CAD/CAM, doing model kit parts in the tooling mockup stage for a model kit simply required the services of highly skilled artisans--sculptors in every sense of the word--no less artists than a Michelangelo…”. The incredible depth and breadth of Art’s contributions to this discussion open the door wide open to the world of model kit creation. The fact that one of the wooden models is of the Bugatti Veyron indicates the sculptor’s arts haven’t been entirely replaced by computer data quite yet! And finally thanx to Bill Engwer for his own hands-on insights and the wonderful photos.

It seems that my original question has been somewhat eclipsed by reality, that reality being that, as Art has explained so well, creating a plastic model kit is a complex and meticulous process taking the creativity, talents and commitments of many people. Identifying a “star” who creates wonderful prototype models which are somehow magically translated into injection molds misrepresents the whole process. Perhaps the role of “kit designer” or team leader is the closest we can come to that – Tom West’s reputation has a lot of that to it. I recall an interview that Hobbylink did with the owner of Model Factory Hiro that made it clear that it was very much the owner of that company’s vision that informed everything they produced. But these people are not necessarily the “hands” that produce the prototypes.

A subset of my original question is the curiosity I have about how the small bits are created, the carburetors, injection pumps, wheels, dashboards and so forth. The level of detail is often mind boggling.

Thanx for all the info, stories and lore provided so far!
B.