Put in reverse: in order to put a classic, extremely rare Duesenberg Model J into gear, metalcasting engineers had to go in reverse.

Jim Schneck wanted in. He had the car--a supercharged 1934 Duesenberg Model J worth more than $500,000--but he wanted the respect of those in the inner circle of Duesenberg collecting. Metalcasting turned out to be his passkey.

Schneck had worked previously with a group of metalcasters on several restorative projects. So, he thought he could count on them to help him through a restoration that would put him in an elite class of collectors, including automotive historian Randy Ema and renowned car buff Jay Leno.

In the past, Schneck had worked with metalcasters to recreate small parts and pieces that were originally cast in low grade "pot" metal. He also had done considerable work on the Auburn supercharged car, recasting the intake manifold and front timing chain case. This time, he wanted to recreate the rare Duesenberg's cylinder head, an expansive project that would eventually produce a 180-1b. (81.6kg) machined casting.

"This is one of the most difficult things that I've ever done," said Schneck. "But there was a need for it, and I wanted to receive the calls from the Jay Lenos of the world. It doesn't have to do with money; it has to do with supplying a need that's been there for 80 years."

Back It Up

The original engine block and cylinder head of the Duesenberg Model J was designed by one of a speed-loving pair of brothers. Drivers first and businessmen second, the brothers--Fred and August Duesenberg--started a company in the early 1920s to make cars that would win races. They launched their line of cars by putting out hotrods like the Model A and X. The now famous Model J made it to the drawing board shortly thereafter, but "Auggie," as he is known among car enthusiasts, was no longer involved in design.

"Part of their racing history made them understand that horsepower was developed in the head," Schneck said. "The idea was new to America. [The Model J] was one of the first production cars with a double overhead cam design. A couple other manufacturers copied it after that."

This revolutionary idea made for some complex castings. It also put a strain on the Duesenbergs' pocketbooks. Before the Model J went to production, the brothers sold their company to E.L. Cord, who had the cash and the means to put together a 180-lb. cylinder head, not to mention the rest of the car's well-crafted components. Cord, who hired Fred Duesenberg as his chief engineer, simply sent the head's design over to Lycoming Engines, a one-stop engine manufacturer that he owned, so that it could be poured in the onsite metalcasting plant. From there, green sand molds were developed, and iron was poured.

"I can't believe they were able to do this back then," said John Humont, Foundry Consultants Inc., Toledo, Ohio, who would work on recasting the head 77 years later. "It absolutely flabbergasted me. They had a horrible time trying to make the thing."

But they did make the thing--nearly 500 times. Between 1928 and 1935, 481 engines were produced. At the end of that stretch, the company had sent all of the Duesenbergs that would ever be produced to the road. The company was driven out of business in the late 1930s, largely for its lack of ability to keep up with other car makers, Schneck said.

Today, the existence of 368 of those original 481 has been confirmed. Around 70 of the automobiles are known to have been destroyed, and 40 have seemingly disappeared from the known record. Some of the remaining 368 have and/or need replica parts to return to working order. And, coming up with replicas can be a chore, especially for complex pieces like the cylinder head.

Time Tested, Failed

Like the Duesenberg brothers' hold on their car company, the metal castings that composed the 481 engines Cord produced were not destined to stand the test of time. At least 10% of the remaining 368 cars need new cylinder heads, according to Schneck.

"Every head that I've seen had repair work done to it," Humont said.

Schneck's own Duesenberg, luckily, contained a properly working cylinder head when he decided to begin the project.

The group Schneck organized, which included himself, Humont, Jim Flanigan of Foundry Consultants, Dave Smith of Cunningham Pattern and Engineering, Columbus, Ind., and Scott Wichlacz of Manitowoc Motor Machine, Manitowoc, Wis., were fortunate enough not to have to reverse engineer the head entirely. Along with the working cylinder head and an inoperable head, they were able to obtain blueprints from the original head. However, the exacting standards of Duesenberg collectors made it necessary to copy the outer appearance of the working head exactly, making the job largely a reverse engineering project.

To engineer the tooling from the existing heads, the prints first had to be consulted to assure that no significant changes had been made to the design when they were poured.

A coordinate measuring machine (CMM) was used on both the sectioned casting and the intact head to show, with extremely tight tolerances, that the heads were similar enough to the prints to serve as a template to work from. The job of rebuilding the original pattern then went to patternmaker Flanigan. He received the full and sectioned castings in November 2004 and had a wood and plastic pattern built by May of the next year.

Revisionist Historians

Duesenberg collectors are a demanding bunch. The reproduced head had to look exactly like the original, or the restoration would have devalued the classic automobile. So, it was important to work backwards from Schneck's head, rather than simply design the pattern and core boxes from the blueprints. Unlike the outer appearances, though, the patternmakers were able to take some liberties with the inner workings of the machine.

"We were able to beef up some of the walls," Flanigan said. "We built the areas up around the spots that broke down on the originals. It took three comparisons back and forth to continually look at the head; it was an intense effort to return the head to as close as possible to the original."

The resulting patterns indeed would produce cosmetics that matched the castings poured at Lycoming, right down to the parting lines. Mong with the cope and drag tooling, Flanigan and one other patternmaker put together 37 different core boxes to make the necessary 45 cores. All of this was done in wood and plastic with careful measuring by hand. In May, it was time to send the tooling on to Smith at Cunningham Pattern.

No Bake, No Breaks

Smith says that some of the tooling had to be modified slightly to make the patterns and core boxes agreeable to metalcasting. But in early 2005, Smith was able to put the tooling to work, after Flanigan and his team had reworked the pattern twice, and Smith and his colleagues performed one rework.

Cunningham Pattern molded the large cope and drag in nobake sand--"it looks like a sarcophagus," said Schneck and constructed the cores primarily with coldbox tooling, though some were also done in nobake.

It was the cores that offered the most difficulty. Before Smith ever received the tooling, Flanigan and his fellow patternmaker went through the pains of understanding how all the cores would fit together to recreate the inner workings of the Duesenberg head.

"The greatest obstacle was having the cores pass over one another," Flanigan said. "It took a long time to figure out which cores went into the others. When you're assembling 45 cores, you have to carefully consider the air and water passages."

As it turned out, the design of those air passages was one of Smith's main concerns, the main reason why the tooling needed some rework to be "casting friendly."

"Anytime you pour, gas is your biggest issue," he said. "If you don't vent properly, all you have is a gassy casting." But after four pours, Cunningham found the right equation. "We took our time, put it together, and drilled the cores. Eventually, it made a nice looking head."

With all of this core manipulation, Smith said nobake sand made the job quite a bit easier than it would have been when it was originally performed in green sand.

"Nobake is more user friendly," he said. "It's harder than green sand. The cores that you have to take in and out--you can't do that with green sand."

And more mold breakage is apt to happen with the long draws associated with cylinder head castings, which Smith calls one of the most difficult pours in the business.

Feels Good, Duesen't It?

The completed casting weighs 220 lbs. (99.8 kg) prior to machining, is 48 in. (121.9 cm) long and represents a significant improvement in metal quality over the original Duesenberg heads. After stress relieving and impregnating the metalwork, Wichlacz machined it down to a still-substantial 180 lbs. (81.6 kg).

The first machined casting was fitted into an ailing Duesey in late August. It made its public debut at the Auburn Cord Duesenberg Festival in early September. The group plans to operate the engine sometime in the months to follow. That will be the final test of the recasting's success, said Schneck.

At first, Schneck anticipates a limited production run of eight to 15 heads. But eventually, he expects the demand, based on response from other collectors, to reach about 50 heads. According to Humont, as time passes, more and more interested Duesenberg owners come forward looking to upgrade their engines. And some of them are interesting Duesenberg owners.

"Jay Leno called," Schneck said soberly. "He's interested in this project."

For More Information

"3-D Laser Digitizing for Reverse Engineering, Mold-making, QA and Rapid Prototyping," G. Hilgers, 1994 AFS Transactions (94-77).

"New Horizon in Nobake Binder Technology," E. Ireland, K. Chang, J. Kroker, 2002 AFS Transactions (02-025).

Shea Gibbs, Assistant Editor