February 3, 2009: On this day, the Duesenberg engine J-370 was ready to go back to meet the car after completing dyno testing. Behind the Model J in today’s lead image are an Alfa Romeo 8C, a Maserati Birdcage, and a McLaren M26… just another day in the shop on Paradise Drive.

I worked on this project through the course of a year at Phil Reilly & Co. and it was my first introduction to the iconic American brand. Owned for a time by movie star Mae West, chassis number 2387 originally had a Murphy Convertible Coupe body and was later redone by Bohman & Schwartz in Pasadena, CA.

Reilly’s had the engine, the gearbox, and the rear axle, while the rest of the car was being finished at Gold Country Auto Restoration in Murphys, CA, in preparation for Pebble Beach.

The Model J Compared to Lincoln and Hispano-Suiza V-12s

The Model J was Duesenberg’s attempt to build the best car in the world without any concessions towards cost. Fred Duesenberg, the engineer behind the company’s designs, chose several features for the engine that were already proven by other makers, such as 4-valves per cylinder and dual overhead cams. But it seems strange that there was no attempt to limit the weight. Both the block and cylinder head were cast in iron, and they surely are in competition for the title of the heaviest engines ever made. The massive crankshaft had an equally massive flywheel, connected via a torque tube to a massive rear axle. (Are you sensing the theme here?)

And yet, when it came time to design the connecting rods, he decided to use a light aluminum alloy! It was recognized as a weak point, because when they built the supercharged SJ model, they were replaced with steel rods.

It is also interesting that Duesenberg chose chains for the cam drive, which is typically seen as a quieter and cheaper option compared to gears (Alfa Romeo 8C, Bugatti Type 57) or shaft drive (Hispano-Suiza H6, Alfa Romeo 6C). Not only chains, but four-row chains. Again with the massive theme. Yet the cam and tappet design choice was known to be a noisy option. Hispano-Suiza went to pushrod-operated valves in their J-12 in an effort to chase silent operation. The J-12 also had aluminum crankcases and cylinder blocks and was vastly lighter than the Duesey J.

Another marketplace competitor was the Lincoln KB. I have built a couple of V-12 Lincolns from that era and they also had aluminum crankcases. The 1932 Lincoln KB could be had for less than $5,000, and custom-bodied versions were in the $7,000 range. Less than the bare chassis price for the Duesenberg! Which was $8,500, or over $200,000 in 2026 money. A 1932 Ford Model B sold for $495, for comparison.

I can make a direct comparison between the Duesenberg and the Lincoln, because I have run both engines on the same dyno in the same year. The Lincoln, a simple, single-cam, L-head design, made more torque at the same RPM, 385 lb/ft at 1,800 RPM, where the Duesey managed 368 lb/ft.

As for horsepower, Duesenberg’s marketing boys were good. It is widely reported that the Model J was good for 265 horses, a wildly optimistic number by some accounts. This engine here produced 215 HP at 4,500 RPM.  They were usually quoted as having a redline at 4,200 and the modern rods may have allowed closer to 5,000… but there’s not another 50 HP in those 500 revs.

Over in Detroit, Lincoln said that the seven liter KB V-12 would output 150 HP, a more humble estimate, and our example here did 193.6 at 3,300 RPM.  So almost 90% of the horsepower of the Duesenberg for a much smaller fraction of the money. And more torque.

The Hispano-Suiza J-12 that I built in 2012 also ran on this same dyno, but sadly I do not have copies of the result. Given that it was eleven liters, I suspect that the torque numbers must have demolished both of these.

For another competing European model, one of the Bugatti Type 57s that we ran on this dyno was close to hitting 300 horsepower, from only three liters, although it was supercharged and running on alcohol. (It also suffered through some teething pains to get there.)

The Model J Engine Rebuild

As it was almost twenty years ago now, my recollection of the job is pretty fuzzy in places. My pictures, from an era before I became more meticulous about getting every detail on disc, show much of the assembly but do not include many of the pieces that I was interested in for this particular article. Such as the main bearings, or the valves, or the pistons and rods! You can never take too many pictures.

The 2008 build sheet says that the crank was nitrided and ground. With the babbitt main bearings, the crank’s main journals can be finished to whatever size is needed to clean up the finish, as the babbitt will be line bored to suit the crank.  This engine replaced the original aluminum rods with a set of Carrillos, engineered to use modern shell bearings, thus the need for nitriding the crank to get a suitably hard finish on the rod journals. The width of the rods requires two pairs of shells mounted side-by-side.

I wrote a little about the crank counterweights and vibration damper earlier this year.

A set of Arias forged pistons with a modern ring package was made, with the final compression ratio coming out at a very low, but period-correct 5.5:1. Imagine what it could do with 8:1!

There’s not much else to say about the bottom end. The oil pump got a thorough service, with its strange cork float mechanism mounted around it that controlled the oil level meter, operated by a tiny set of bronze bevel gears. I forgot until looking at the photos that it also had a pressure relief valve with a flexible cable for adjustment, and a third cable that passes through the pump body, for a remote oil drain valve.

The Reproduction Cylinder Head

The head that came with the project was deemed to be scrap, for reasons I don’t recall, more than likely excessive cracks. New ones had recently become available and in general, the replacement head was a beautiful casting with quality machine work. I did a little clean up of the ports where the bored section had a rough transition to the runner walls, but that was about it for the initial setup.

It is worth noting that the intake port design of the Model J was… how shall I describe it? Not as direct as the exhaust. Each pair of exhaust valves got a single port, so eight ports for eight cylinders on the right side of the engine. The intake ports were siamesed, with only four ports for the eight cylinders. The air flow had a quite sharp turn to make once entering the head. When you also consider that the front and rear ports were much farther away from the centrally mounted carburetor, it was less than ideal in terms of airflow and mixture distribution.

It is also fair to say that the engines from competing manufacturers were not too different in regards to the carburetor location, and the Hispano-Suiza also had siamesed intake ports. The Lincoln was weirder, with each bank’s outer two cylinder pairs using siamesed intakes, but the center pair had individual ports. The Lincoln’s exhaust manifolds had six ports, one per cylinder. So each block had ten ports.

The new head would get new tappets with cams refinished by Megacycle, but we used the removable tappet guides from the old head, along with most of the studs. Late into the engine assembly process, when the bottom end was already completed, I started to test fit everything into the new head. But there was a major problem…

When the hold-down clamps for the tappet guides were tightened, the tappet guide bores distorted and the tappets would seize. It did not matter what the torque number was… if you backed them off until the tappets were free, the clamps were far too loose to do the job.

Each pair of tappet guides is secured into the head with two intricate little hold-down clamps, steel forgings that have three parallel machined surfaces. The top surface seats against the nut that pulls the part down over a stud, the two feet that press against the shoulder of the tappet guides, and the underside of the stud hole, which is higher than the feet and not intended to touch the head.

This issue caused much head scratching in the machine shop. Where was the problem? How was this distortion generated? Bolting the parts into the old head to test the issue did not produce the same error. Many things were measured and checked. And though I was not privy to all of the discussions, I did eventually assemble the head with the solution that they arrived at.

Each pair of clamps got its own numbered pair of spacers, ground to a height that made the two feet press sufficiently tightly onto the guide shoulders, when the nut was torqued to stretch the stud and keep it from coming loose. So instead of the bottom of the clamp around the stud being free in the air, with the two legs of the clamp flexing under tension, the whole package became one with the head casting.

This is a perfect example of how a planned assembly schedule can go straight out the window! Resolving this unanticipated issue took some time.

The top end eventually came together without further drama, although the valve lash adjustment on the Duesey was unique to me. It used a stack of very thin shims in an adjustment cap on top of the valve stem, and this is one of the places where I can’t believe I didn’t take pictures of these parts. The two-piece adjuster cap is threaded so that it can be tightened together with the shims inside it. Why didn’t they use the thread as the adjusting mechanism, since they were already investing in a threading operation? Why not a single piece stem cap that is ground to size like everyone else? Fred, what were you up to?

The engine left Reilly’s on February 4, 2009 and the gang at Gold Country got the car onto the green at Pebble in August of that year. It really is an amazing car. I was at the event to support the Alfa Romeo TZ2 that we also finished that year, and it was nice to be able to see the Duesey on the green.

And for a parting shot, I found this wonderful picture at duesenbergs.com while researching for this article, reputed to be this car in its first body configuration. While trying to chase down the image source, I was able to locate the building in the picture, in Spokane, WA, and it still exists, with essentially the same design! How cool is that?

What Does it All Mean?

Legendary cars like the Duesenberg Model J all have their own mythology and dedicated followings. Having been inside the engines of this car and of some of the competing brands, I have my own opinions on the myths and legends. Personally I think the Hispano-Suiza is the superior machine. But the brute force that the Lincoln is capable of is also very impressive, with mass production keeping the cost in check.

And then there are the Bugattis, like the 1931 Type 50 that was featured in an earlier article – 200 horsepower from a supercharged five-liter straight-eight.

It could also be that my body harbors a grudge towards the Duesenberg, because shifting those heavy iron castings around left some lasting marks.

Do you have a favorite, or an argument for your preferred model’s superiority? Let us know in the comments below. And subscribe to get notified when the next One Day feature comes online.