Main Article

Automotive engine design and analyses has changed dramatically and is vastly improved since the Ford Model A engine was designed and analyzed in 1927. Have you ever wondered why even the best rebuilt or highly modified Model A engine has a useful life that is just a small fraction of the useful life of a modern engine? This article will attempt to answer that question and present an engineering design study that demonstrates what can be accomplished by substituting four redesigned parts into a Model A engine. By substituting these four redesigned parts, a stock appearing Model A engine can have the reliability and longer life of a modern engine, and a hot-rodded engine will have a much higher probability of staying together. Readers of this article will also learn about modern engineering methodology, understand the reasoning behind engineering design decisions, and learn how a collection of sand cores can come together to form the cavities of a complex casting. For additional information, readers are encouraged to do Internet searches on the words, phrases, and terminology used in this article.  This article presents a summary of what has been accomplished. And lastly, this article has been written to determine if there is enough interest for this engineering study to continue and become real hardware. I apologize for the length of this article, but there is a vast amount of information to present. 


DIFFERENCES BETWEEN MODEL A AND MODERN DESIGN PRACTICES

If compared to a modern four-cylinder engine design, the Model A engine differs and is deficient from modern design practices in several areas as itemized below:

EVOLUTION OF THE MODEL A FORD ENGINE


The Model A engine and its derivatives were manufactured beginning in the fall of 1927 and continued until 1958. During this manufacturing span, several improvements were made to the basic design, however all of these improvements still fall short of modern engine design practices. The following chart illustrates the evolution of the Model A engine family and shows many of the changes that were made throughout its evolution. 


Evolution

Comparing the evolutionary chart above to the design differences noted earlier, it can be seen that most of the design deficiencies between the Model A engine and a modern engine were narrowed throughout the evolution. 

Also, as can be seen from the evolutionary chart, the latest engine, the G28T engine was a vast improvement over the Model A engine; however it is still not up to modern engineering standards since it only had three main bearings and a partially counterweighted crankshaft. It should also be noted that the exterior appearance of the G28T engine looked like a modified Model B engine, which is somewhat different than a Model A engine. 

RACECARS


Throughout the 1930’s, racecars powered with Ford Model A and Model B engines were the crowd-pleasing underdogs in big car racing.  The Fords would easily out-accelerate the competition of expensive pure bred (Miller and Offenhauser) equipped cars at the beginning of the race and also coming out of the corners in oval track racing, however they would often break before the end of the race. Aftermarket heads were available to fit any budget and included flathead, F head, I head, SOHC, and DOHC configurations. They were normally aspirated, and the more potent Fords with compression ratios exceeding 12:1 could produce over 250 HP from 200 in*3. Tetra-ethyl lead and other chemicals were readily available and freely added to fuel to boost performance. HAL heads and others were cross flow and featured hemispherical combustion chambers long before Chrysler introduced its’ famous hemi in 1951. The 1930’s were an era of creativity where good mechanics would often beat engineers in discovering how to obtain greater power from engines. Fronty even made a DOHC configuration that was called a "stagger valve" which featured four valves per cylinder, where diagonally opposite valves were either intake or exhaust. This configuration required intake and exhaust manifolds on both sides of the cylinder head. 

The Fords were fast off the line and out of the corner because their bottom end (connecting rods and crankshaft) was extremely light (low inertia) and strong. However, they were also extremely fragile because that same light bottom end construction could not sustain high loads for any length of time. It was common to install a new crankshaft after a few races to avoid breakage. Crankshaft deflections and stresses were excessive and far exceeded modern engineering limits regarding fatigue. The fatigue properties of materials were not well understood. Towards the end of the 1930’s, many of the same manufacturers that originally supplied the more potent overhead valve conversions also supplied 5 main bearing crankshafts with counterweights and matching new cylinder blocks to increase reliability (HAL, Dryer, Cragar). Model A and B based engines were expensive to maintain but were dominant (at least in numbers) in big car racing until WW2. 

After WW2, racing took a new direction. Veterans returning from the war came home with fresh ideas and a desire to participate in racing instead of just being spectators. The interest and participation in midgets, drag racers, and jalopies exploded. Ford flathead V8’s became dominant due to a variety of reasons including engineered counterweights, pressure lubrication, stiffer crankshaft, insert bearings, economical to maintain (no need to change crankshafts to avoid fatigue failures), and the huge variety of inexpensive aftermarket speed equipment that was available.


COMMENTS ON MODIFIED ENGINES

It seems that almost everyone involved with rebuilding a Model A engine today has some opinion as to how to improve it in an attempt to reduce the design deficiencies listed above. Unfortunately, there is no real engineering behind the majority of these improvements, and many of them are poorly implemented. 

Counterweights are added to crankshafts without any calculation as to how large they should be or where they should be located, various odd-ball bearing inserts from tractors, generators, and small foreign engines are implemented without any concern for bearing loads, tang location, and availability in the future. Crankshaft thrust is often moved from the rear main to the center main. The rear main wall of a Model A cylinder block is 5/16 inch thick and reinforced with ribs while the center main wall is 5/32 inch thick with no ribs. Moving the thrust from rear to center not only overstresses the center main wall of the cylinder block but also introduces additional bending stress in the rear half of the crankshaft from clutch release loads. Crankshafts are failing from fatigue because they have had their rear main seal area diameter increased by welding for adaptation of a seal from some other engine without concern for heat treatment or flywheel mounting flange warpage. Crankshaft flywheel flange runout in many of today’s modified engines is excessive and this only hastens fatigue failures. 

Holes are drilled in various places in the cylinder block for the routing of copper lines that eventually work harden and crack. Aluminum has been selected for many aftermarket parts that interface with original cast iron or steel parts without any consideration that aluminum expands three times as much as iron or steel for the same temperature change, or that the stiffness of aluminum is one third the stiffness of steel. Camshafts are notched to clear aftermarket connecting rods and some have even been relocated outboard in the cylinder block in an effort to tighten the mesh between worn camshaft and oil pump drive gears. 

Most of today’s modified engines are one of a kind, and have lost their ability to interchange parts with a stock engine. In addition, many of these "improvements" can actually cause trouble in later rebuilds if the location of original machined surfaces and datums has been altered from stock. 

And sadly, many modified parts have been delegated to the scrap pile on subsequent rebuilds if the rebuilder is trying to return the engine to stock, cannot determine what was done in the past, or has a different idea of what and how modifications should be implemented.


GROUNDWORK FOR REDISIGN

One of the reasons that modern engines have longer life and greater reliability is that their crankshafts are stiffer, are better supported, and fully counterweighted. Compared to a Model A engine crankshaft, the crankshaft in a modern four cylinder engine has larger diameter shorter bearings, a main bearing on each side of every cylinder (5 mains), and a fully counterweighted crankshaft with two equal counterweights for each cylinder. This engineering study and redesign features a 5 main crankshaft, larger diameter short bearings, and full counterweights.

Another reason is that with modern engineering, parts in dynamic environments can be optimized through the use of Solid Modeling and Finite Element Analysis 

CONSTRAINTS IMPOSED

To maintain the original appearance of a Model A engine and the interchangeability of parts, the following three constraints have been imposed on this engineering study and redesign. These constraints are: 


In other words, an engine assembled with the four redesigned parts of this engineering study with all other parts stock will have an exterior appearance identical to a Model A engine but have the life and reliability approaching that of a modern engine. If you haven’t guessed it by now, the four parts that have been redesigned in this engineering study are the crankshaft (A-6303), connecting rod (A-6200), cylinder block (A-6015), and oil pump drive bearing (A-6560).


ENGINEERING METHODOLIGY (OLD VS. NEW)

One of the reasons engine design has improved dramatically over the years is because of the advances in engineering. Up until a few years ago, a design engineer would create a new design with orthographic views on a

two-dimensional drawing, and from that, a mass properties engineer would compute weights, centers of gravities and moments of inertia, and then another engineer would perform a stress analyses at the points he considered critical. Materials were selected to endure the stress based on their yield strength and a factor of safety. The factor of safety was an empirical number (educated guess) derived over years of experience, and varied depending on whether the stress was constant or cyclic, and what the implications would be in the event of failure. After release from engineering, the two-dimensional drawing that completely described the part would be given to pattern makers and machinists would then make the part. For parts that were produced in production quantities, CNC machines were programmed from the drawing. Testing was an important part of the engineering and manufacturing process as it would validate a new design and many times find a weak point that was overlooked or an error that occurred during design, analyses, pattern making, or machining. Results from testing were reported to the design engineers and they would make changes for the next iteration. In a complex assembly, every part is either over-designed or under-designed. This method of engineering and manufacturing is adequate for static design and will continue, but it falls short of modern engineering standards for dynamic design.

 Today’s modern engineering for dynamic design is completely different from anything in the past and is accomplished by solid modeling using computer programs such as SolidWorks, ProE, and Ideas. A solid model is not something you can hold in your hand but instead is a three dimensional image generated on a computer. In modern engineering, the solid model describes the part instead of the drawing. Today’s drawings are used to supplement the solid model by calling out notes, surface finishes and other things like Geometric Dimensioning and Tolerancing (GD&T). Today, a design engineer creates a solid model that can then be analyzed for stress and strain using Finite Element Analyses (FEA). FEA works by dividing the solid model into a finite number of very small elements. Each of these individual elements is computer analyzed for stress and strain. Modern engineering consists of several iterations of solid modeling followed by FEA to optimize a design. Materials used in dynamic applications are selected based on their fatigue strength, which varies with the number of cycles of stress. Fatigue strength of a material is shown on a S-N curve where S is stress (vertical axis) and N is the number of cycles (horizontal axis). Parts that have stresses and number of cycles to the right of the curve will eventually fail (when the number of stress cycles is reached). Parts that have stresses and number of cycles to the left of the curve will have infinite life.  The Fatigue Endurance Limit of a material is that level of stress where the material will never fail even with an infinite number of cycles. Fatigue properties of materials were not well understood until the space race. Launch vibrations (many load cycles in a very short time) would cause failures, and engineers realized that the strength of a material is also dependent on the cumulative number of cycles of stress. Testing today is not as important as in the past, because of the improvements in analyses and the selection of materials based on their S–N curve.

Solid models are also machine shop friendly. Modern pattern makers and machinists easily convert the solid models into instructions for CNC machines. The chance for errors is minimized. Casting wall thickness variations and core shifting is a thing of the past with CNC machined core boxes.  Once a CNC machine is set up, every part produced is identical. 

SolidWorks 2004 was used for this design study.


Figure 1, Exhaust Port Core

Figure 1 is a computer generated solid model of the exhaust port core for the redesigned Model A engine cylinder block. This is the part that a foundry would create as a sand core, and then pour gray iron around it. After the iron solidified, the sand core would be broken out and the resulting cavity would become an exhaust port cavity in the cylinder block. 

WISH LIST FOR REDESIGN

The following is a wish list of changes that have been considered in the upgrade of the four redesigned parts to modern engineering standards.

OPERATING CONDITIONS ASSUMPTION

 In addition to the wish list of changes, operating conditions need to be assumed for this engineering study. These operating conditions are required to perform an FEA (finite element analyses) that will determine the stress and strain at every point in each of the parts. Knowing the stress and strain enables the engineer to choose appropriatmaterials, heat treatment, and surface finish for that part. 

For this engineering study and redesign, the condition chosen for continuous operation (unlimited fatigue life) is 150 horsepower at 5000 RPM (158 lb-ft @ 5000 RPM). These numbers equate to about 112 miles per hour with a 3.78 rear end ratio with 4.75x19 tires. Although this condition appears unrealistic, it has been assumed for calculating stresses in this engineering study, and will provide a margin of safety for loads that can’t be reliably calculated such as shock loading, loads due to detonation, loads due to poor machine work in a later rebuild, etc. 

SUMMARY OF WHAT CAN BE DONE

To make a long story short, every item on the wish list can be accomplished with this redesign and they all can be incorporated within the design constraints listed earlier. In other words, a modern engine can be contained within a cylinder block that looks exactly like a Model A engine from the outside. Read on for a description of the redesigned parts and how they were designed. 

 ENGINEERING STARTING POINT

 Everything has a starting point. The starting point for this design study was a copy of an original Model A drawing (A-6015) of the cylinder block dated March 28, 1929.  This drawing contained all dimensional information including tolerances needed to cast and machine a complete cylinder block. Additional information was obtained from copies of original Model A drawings for the connecting rod and crankshaft, inspection and measurement of several original cylinder blocks, and by measuring the intake port and water inlet baffle of a Model B cylinder block 

CONNECTING ROD DESIGN (A-6200)

 A design study was initiated and several engineering design layouts were performed to optimize the new connecting rod profile. There were two objectives for this initial design study. The first was to provide the largest possible big end bearing on the connecting rod while maintaining stock crankshaft stroke (4.250 inch); stock connecting rod length (7.500 inch) and not having the connecting rod big end interfere with the camshaft, inner walls of the crankcase, or oil pan. The second objective of this initial study was to maximize the width of the connecting rod column (cross section between big and small ends) to maximize column strength. 

From this initial study and layouts, an accurate full-scale model with moving components was constructed out of acrylic plastic to dynamically demonstrate validity of the profile.

 Figure 2 is a photograph showing the acrylic model. Although it appears that the cam lobe will hit the connecting rod in the model, in reality the lobes are fore and aft of the connecting rod, and the only clearance issue is with the base circle of the camshaft. 

Figure 2, Acrylic Full Size Model 

 The new connecting rod big end bearing is 2.000 inch diameter and the width of the big end of the connecting rod is 1.120 inch. The bearing insert is a Federal Mogul part # 2020CP that has been used in a connecting rod application in a variety of General Motor’s 4, 6, and 8 cylinder engines produced from 1955 until 2002. Replacement inserts are readily available, economical, available in many under sizes, available in many configurations depending on application, and should be available for many years to come. This bearing insert is also available in heavy-duty configurations for severe loading (Clevite 77, basic part # CB-745), and from many other manufacturers.

 One of the original applications for this bearing is in a small block Chevrolet V8. Bearing insert loads in the new connecting rod of this study are similar to loads found in a small block V8 engine producing 300 HP at 5000 RPM. This re-engineered Model A engine has being engineered for 150 HP at the same 5000 RPM with half as many cylinders, so horsepower per cylinder is identical.

 The connecting rod big end of this re-engineered engine is split on an angle of 76 degrees from vertical. This angular split was necessary to accommodate the 2.000 in. diameter big end bearing while maintaining clearance between the big end profile, camshaft, and cylinder walls. 

Precision studs with dowel pin accuracy are threaded into the upper rod half to provide accurate cap alignment. Both ends of these studs are accessible on assembly, so a micrometer can be used to measure stretch, which is the preferred method of measuring fastener preload. The cap is attached with a pair of MS21042-6 nuts. The width of the upper end of the new connecting is equivalent to a stock rod and accepts the same pair of wrist pin bushings (Federal Mogul 8229X) found in a stock connecting rod. 

Figure 3, Connecting Rod 

Figure 3 is a computer generated solid model of the redesigned connecting rod. 

Connecting rods are subject to tension and compression loads. The compressive load is gradually applied as the fuel burns and occurs over a period of time (power stroke), however the tension load is very severe and sudden (like an impact). In a normally aspirated engine, peak tension loads are usually much greater than compression loads. That is why race engine mechanics often speak of connecting rods stretching. Tension loads are due to piston acceleration.   

Maximum connecting rod tension loading occurs at TDC at the end of the exhaust stroke and beginning of intake stroke when the connecting rod has to stop and then change the direction of the piston (Crank angle at 0 or 360 degrees). This is the load that stretches the connecting rod, causes the big end to become oval, and pulls at the fasteners. The acceleration load G expressed in g’s can be approximated at any crank angle from the following equation, which can be found in many handbooks. 


G = (((RPM*2 x S) / 2189) x (Cos a + ((r/l) x Cos 2a))) / 32.174


RPM is revolutions per minute,

S is the stroke (4.250 inches) 

2189 is a constant to get the units correct 

a is the crankshaft angle (0 or 360 degrees at TDC and 180 degrees at BDC)

Cos a is the cosine of the crank angle and varies between 0 and +/-1 throughout the 360 deg of crank rotation 

r is the crank throw (2.125 inches) 

l is the length of the connecting rod (7.500 inches)

32.174 is the acceleration due to gravity (ft/sec*2)


 When solving the above equation, positive results (+ g’s) will cause tension loads and negative results (-g’s) will cause compression loads in the connecting rod.

 Solving the equation at TDC (0 or 360 degrees where cos a = 1, and cos 2a = 1) and 5000 RPM, the acceleration G is +1936 g’s.

 Solving the equation at BDC (180 degrees where cos a = -1, and cos 2a = 1) and 5000 RPM, the acceleration G is -1081 g’s  (56% of TDC acceleration)


 For stress calculation purposes (and a sample calculation), assume that the connecting rod is going to fracture right across the middle of the wrist pin. The area of the new connecting rod at this fracture plane is .4646 in*2 and has to resist the g loading resulting from the weight of the piston, rings, wrist pin, and upper piece of connecting rod that are all attempting to fracture away. For calculation purposes, the weight of all parts attempting to fracture away is assumed to be 2.00 lbs. (Actual weight of a stock piston, wristpin, and rings is 1.55 lbs.). This worst-case assumption compensates for loads that cannot be accurately calculated such as shock loading, loads resulting from inaccuracies in later rebuilds, and increased loading if a heavier than stock piston is used. 

With the above assumptions, the force attempting to fracture the small end of the connecting rod at TDC due to cceleration at 5000 RPM is 1936 x 2.00 lbs = 3872 lbs tension. This 3872 lb load is acting on the .4646 in*2 cross sectional area. Stress in the connecting rod at this location is 3872 lbs / .4646 in*2 = 8334 lb/in*2. Therefore, a material with a fatigue endurance limit greater than 8334 lb/in*2 is required. 

Likewise, the tensile force at the big end of the connecting rod at TDC can be calculated. Here the load is much greater because the part attempting to fracture away also includes the weight of the upper part of the connecting rod. The upper part of the connecting rod weighs 1.136 lbs, so the total load is (2.00 lbs. + 1.136 lbs.) x 1936  = 6071 lbs. This is the load that the two fasteners attaching the cap must endure.

 Similarly, the g load, stress, and strain at every crank angle and RPM can be calculated at every location on the connecting rod with FEA. This is all done on a computer. FEA results can be displayed on the part being analyzed where different colors represent different amounts of stress or strain. An FEA display looks similar to a weather chart where different temperatures are represented as different colors.

 Have you ever wondered where the bearing tang notches should be? Modern engine design places them on the same side of the journal to insure that there is no binding in the event that the parting line is not exactly centered on the bore. But what side? Maximum loading occurs at TDC which means the bearing insert half in the connecting rod cap reacts the highest load while at this very instance, the upper bearing insert sees no load. This combined with the direction of rotation of the journal dictates that the notches should be placed on the side closest to valves. 

CRANKSHAFT DESIGN (A-6303)

 The new crankshaft has five 2.000 inch diameter main bearing journals. Crankshaft thrust is controlled at the rear main bearing. Generous fillets are used throughout to mitigate stress risers.

 The new crankshaft has four .1875-inch diameter drilled oil passages from the front, center, and rear main bearing journals that supply pressurized oil to the connecting rod journals. 

Counterweight calculation is an engineering determination of how large the counterweights should be. Counterweight calculation is different than balancing. Balancing is making things equal. A stock Model A crankshaft (without counterweights) is balanced when it’s center of gravity lies along the centerline axis. Small amounts of weight are added or removed to get the center of gravity on axis. Balancing can be either static (pistons, wrist pins, and connecting rods) or dynamic in the case of things that rotate (crankshaft, flywheel, etc).

 Stock Model A and 1st generation Model B crankshafts have no counterweights. The 2nd generation Model B crankshaft had integral counterweights equivalent to about 40% of modern engine standards, and the late ‘30’s factory rebuilt Model B engines had pressed on counterweights equivalent to about 60% of modern engine standards. In both cases, the counterweights are placed on one side of the connecting rod instead of being equal on both sides of the connecting rod as on a modern engine. Placing the counterweight on one side introduces cyclic bending loads into the crankshaft (deflection and fatigue) with every revolution. These cyclic loads increase as the square of the RPM. Each doubling of RPM causes loads (stresses) and deflections to increase four times. 

The new crankshaft resulting from this engineering study has a pair of identical counterweights for each cylinder that are spaced equally fore and aft from the connecting rod as on a modern engine. These counterweights take care of what engineers call "first order effects" as they reduce crankshaft stress and deflection. 

Figure 4, Counterweights and Crank Arm 

Figure 4 is an exploded view that illustrates the pair of identical counterweights on the left and the crank arm on the right that is used for each cylinder.


The equation used to determine the amount of weight for the pair of counterweights is shown below.

CW x A = (CA x B) + (BE + .5 x SE) x C where:

CW is the weight of the pair of counterweights

A is the distance from centerline axis of crankshaft to cg of counterweights

CA is the weight of the crank arm

B is the distance from centerline axis of crankshaft to cg of crank arm

BE is the big end weight of the connecting rod

.5 is a constant (good up to ~ 7000 RPM, increases to .55 @ 18,000RPM)

SE is the small end weight of the connecting rod including weight of piston, wrist pin, and rings

C is the distance from centerline axis of crankshaft to centerline of connecting rod journal (2.125 inches), (1/2 of 4.25 in. stroke)


 There was just enough room in the crankcase for the 100% counterweights if they are made from the same material as the crankshaft (ductile iron). The counterweights could become less if the crank arm, connecting rod, or piston were lighter. The counterweights could also become smaller and have the same effect if they were made of a denser material such as tungsten.

Second order vibration effects are another problem.  Modern inline four cylinder engines incorporate two balance shafts with counterweights turning in opposite directions and at twice engine RPM. These balance shafts cancel a second order vibration caused because the two ascending pistons and two descending pistons do not always have identical opposing acceleration. Remember the connecting rod accelerations calculated earlier at TDC and BDC (+1936 g’s and –1081 g’s).  Counterweights reduce crankshaft deflections and stress, but balance shafts do not. Balance shafts reduce loads on items external to the engine and increase passenger comfort. Since balance shafts do not reduce stresses in internal engine components and since there is no physical room for them in a Model A crankcase, they have not been considered in this engineering study.

The forward end of the new crankshaft is identical to the forward end of a stock Model A crankshaft so all of the interfacing parts (key, gear, slinger, pulley, ratchet nut) fit without modification.

 At the rear of the new crankshaft, the flywheel interface and location is identical to a stock Model A, however the flange is much thicker. This thicker flange is required to provide a sealing surface for a pair of radial lip seals described later and an added benefit is that it shortens the length of the rear main journal, which results in a stiffer crankshaft.

 Figure 5 is an illustration that shows the new crankshaft from the rear. The flywheel mounting holes and dowel pins are not shown. 

Figure 5, Crankshaft

The weight of a stock Model A crankshaft is 28 lbs. The 2nd design Model B crankshaft with integral (40%) counterweights weighs 47 lbs., and the late ‘30’s factory rebuilt Model B crankshaft with pressed on (60%) counterweights weighs 60 lbs. A crankshaft from a 176 in*3 Offenhauser engine that produced 460 HP in 1950 weighed 67 lbs. (Walton, Offy, pg 121). The new design crankshaft with 100% counterweights made from the same material (ductile iron) weighs 90 lbs. If the counterweights contained tungsten (added cost), overall crankshaft weight would drop substantially.


 MAIN BEARING CAPS AND REAR MAIN SEAL DESIGN

 Five main bearing caps are utilized in this redesign. The front and center caps are identical to each other, and the two intermediate caps (between cylinders 1&2, and 3&4) are also identical. 

The front, center, and rear main each use a pair of bearing inserts. The two intermediate main bearings (between cylinders 1 & 2, and 3 & 4) each use a single insert. All 8 inserts are identical, and are the same bearing inserts used in the connecting rods (Federal Mogul 2020CP, twelve sets required per engine).

 All main bearing insert tang notches are located on the side of the engine away from the valves because the upper main bearing inserts are loaded heaviest when they react peak connecting rod tensile loads. 

Figure 6, Rear Main Cap 

Similar to the design of the new connecting rod cap, the new rear main cap shown in Figure 6 is indexed and fastened with a pair of precision studs threaded into the cylinder block. An o-ring gland is provided around each attachment stud hole to prevent an oil leak. Crankshaft thrust is controlled with a pair of identical bronze thrust washer halves attached with three brass flat head screws (screws not shown). An identical pair these thrust washers is found in the new cylinder block. 

A pair of 5 inch OD x 4 inch ID x 1/2 inch wide radial lip oil seals (National #415035) are used at the rear and seal against the new wider crankshaft flange. A matching recess is provided in the new cylinder block to hold these double seals. 

The new rear main cap assembly has the insert bearing halves positioned with a .286 inch gap. This gap between inserts lines up with the drilled oil passage in the crankshaft that provides oil to connecting rod journal 4. This view of the rear main cap also shows the groove for the stock cork pan gasket that is identical in location and dimensions to a stock Model A rear main cap. 

Figure 7, Lower View of Rear Main Cap 

Located just below the bronze half thrust washer, this view (Figure 7) of the rear main cap shows the drilled passage that vents the rear seal cavity to the crankcase. Also seen at the entrance to this drilled passage is a vertical ledge. When the rearmost crankshaft counterweight passes by that ledge, windage from the counterweight creates a slight vacuum (venturi effect) to help evacuate the seal cavity. It works by using the same principle as the crankcase vent (draft tube) on a pre-smog automobile. The pre-smog draft tube extended to below the engine and was typically cut off at a 45-degree angle. As the car moved, air passing by the draft tube created a very slight vacuum to evacuate the crankcase (venturi effect). 

Figure 8, Front and Center Main Cap 

The front and center main caps shown in Figure 8 are identical, and include a pair of bearing insert halves with an .188-inch gap between inserts. This gap lines up with the drilled oil passages in the crankshaft that provide oil to connecting rod journals 1, 2, and 3. Note the ledges on the sides of the cap that will index with grooves in the cylinder block to keep the cap aligned. This feature allows the attachment bolts to be loaded in pure tension without having to react bending or shear loads. 

Figure 9, Intermediate Main Cap 

The two intermediate main caps illustrated in Figure 9 that are used between cylinders 1 & 2, and 3 & 4 are identical and include a single bearing insert. Again, note the ledges on the sides of the cap that will index with grooves in the cylinder block to keep the cap aligned. 

CYLINDER BLOCK DESIGN (A-6015)

 Unlike the other engineering and illustrations in this article that show parts, the engineering and illustrations of the new cylinder block show voids. The reason for this is to make the engineering easy to implement at the foundry. If you were to send a drawing (or solid model) of the cylinder block to a foundry, the first thing that the foundry engineers would have to do is design the core boxes that are needed to create the sand core for each of the cavities. In the case of this new cylinder block, seventeen uniquely different cores are required. The other advantage to modeling the voids instead of the actual part is that undercuts and other unsuitable geometry that would complicate core box construction can be avoided. In this design study, each cavity was modeled separately and then combined into an assembly. The assembly of solid models was then checked for wall thickness and interference, and adjacent models (cavities) were checked for compatibility (feature location). 

Figure 11, Assembly of Crankcase Cores, Left Side 

Figure 10, Assembly of Crankcase Cores, Right Side 

Figures 10 and 11 above illustrate the assembly of solid model crankcase cores. The exhaust port cores are red, the intake port cores are blue, the water jacket core is green, the valve chamber and timing gear cores are gray, and the cylinder and crankcase cores are brown. These cores will become the hollow areas of the new cylinder block.

 All cores of the new cylinder block are designed to provide 150 inch of extra material at each of the machined surfaces. This extra material is called "machine allowance" and is the nominal amount of material that is removed at the machine shop. 

The extra protrusions shown are called core prints. A core print is an extension of a core that will fit into a corresponding recess of an adjacent core that will both position and hold each core in proper orientation. 

The water jacket core is comprised of two parts that come together on an interface that follows the swept centerline of the valve ports. This split is necessary for two reasons. The first is to provide clearance for assembly of the valve port cores, and the second is to simplify water jacket core box construction. 

The water jacket core in the new cylinder block is identical to the water jacket core found in a stock engine with the following exceptions. The intake passage contour has been changed to follow the new streamlined intake port cavity described below, and the exterior has been made smaller by 1/32 inch to accommodate the 3/16-inch exterior wall described below. Figure 12 shows the upper and lower water jacket cores. 

Figure 12, Water Jacket Upper and Lower Cores 

Figure 13 shows four of the five exterior cores that will determine the shape of the exterior of the cylinder block. The top core is not shown. They match the exterior of the cylinder block described on the original Model A drawing with the following exceptions.

 The original drawing did not show parting lines. Parting lines are where foundry cores come together and usually result in a thin fin of excess material on a casting that must be ground off. Parting line location was determined by measuring several original cylinder blocks. 

When comparing the drawing to several original cylinder blocks, it was noted that several fillet radii and other small features did not match. These items were evidently left to the discretion of the original pattern makers. The solid model created for this study is a combination of the drawing and these deviations. It should be noted that an original cylinder block could be laser scanned to create an exact solid model. This is discussed later on under "Status of Engineering". 

Figure 13, Exterior Cores 

Three minor changes occurred to the exterior of the cylinder block during the spring of 1929. The length of the serial number boss was increased to 3 ¼ inches as the 1 millionth Model A was produced, the two bosses at the rear of the cylinder block used for throttle mechanism attachment were changed to one long boss, and a small bump was added to the exterior of the rear cam bearing to provide wall thickness for the new oil hole feeding the camshaft. There are no other known exterior changes through the end of production. The new cylinder block described in this engineering study features these changes. By changing the exterior cores shown in Figure 13, cylinder blocks for earlier configurations are possible.

 In a stock Model A engine, all wall thickness is 5/32 inch with the exception of the rear main wall, which was thickened twice during production to a final nominal thickness of 5/16 inch and the cylinder walls which are ¼ inch. In the new design cylinder block, all wall thickness is 5/32 inch with the following exceptions. The exterior water jacket wall and the main bearing webs all have 3/16 inch walls, the rear main bearing wall is 5/16 inch, and the cylinder walls are ¼ inch. The new design cylinder block also has larger ribs at the rear wall for added strength.

The front main bearing web on a stock Model A engine is close to the rear of the bearing.

This new design cylinder block has the web centered on the main bearing for added strength. The following cross sections, Figures 14 and 15 with various cuts show many of these walls. 

Figure 14, Cross Section 

Figure 15, Cross Section 

The intake ports and valves in a stock Model A engine are 1 3/8 inch diameter and have a 5/32 inch wall to water jacket. Although the intake valves in a Model B engine are the same 1 3/8 inch size, the intake ports are 1 ½ inch diameter with the same 5/32 inch wall to water jacket. Attempting to enlarge or straighten the ports on a stock Model A engine is not advised since the water jacket becomes very thin. 

The new cylinder block maintains all stock interfaces and basic dimensions in regards to the valve train. Nothing is changed from a Model A engine except as follows. The intake ports of the new cylinder block are the same larger size as a Model B engine, and the intake passages of the new cylinder block have been streamlined and straightened by reducing the total number of bends. In the new design cylinder block, the intake charge flows into the intake port, and then makes an immediate horizontal 45 degree turn either left or right, and then a 90 degree vertical turn to get to the underside of the intake valve (total angular change is 135 degrees). The intake port in a stock Model A or B engine continues straight for a short distance before making the turn towards the valve (total angular change is greater than 135 degrees). Figure 16 shows the new cylinder block intake port core with its’ streamlined passage. 

Figure 16, Intake Port Core 

The exhaust valve seats in the new design cylinder block will be hardened steel. 

The new cylinder block has five cam bushings. The cast iron bosses in the new cylinder block that support the camshaft have been enlarged from stock to provide room for the replaceable bronze bushings. Each of these bushings has a 1.562 inch ID and a 2.000 inch OD. This wall thickness at the rear bushing provides enough material for an o-ring groove and face seal at the interface between the new cylinder block and the flywheel housing. As a result, the flywheel housing gasket (A-6396) and flywheel housing to cylinder shims (A-6400) are no longer required. 

The center cam bushing is sealed at both the fore and aft ends with radial lip seals (National #473694) to minimize leakage and maintain oil pressure. 

The Model A engine has a primitive lubrication system. Oil is pumped from the oil pan to the valve chamber and from there, drains to the main bearings by gravity. Excess oil pumped into the valve chamber runs down the oil return pipe assembly (A-6645) where it flows into a tray (A-6688). From this tray, dippers on the bottom of the connecting rods scoop oil for lubrication. On startup, Model A main bearings suffer from a lack of lubrication because all of the oil in the valve chamber has drained back to the oil pan by gravity, and it takes a short period of time for the oil pump to replenish the oil in the valve chamber. The Model B engine has a better system with low pressure to the mains, but still has dippers on the connecting rods. The G28T engine is greatly improved over Model A and B engines by having a full oil pressure system with a drilled crankshaft so that all bearings received pressurized oil. 

Preliminary oil galley and feed line routing for the new cylinder block was worked out on paper, and then made into an acrylic full-scale model to 3-dimensionally demonstrate routing and clearances. A photograph of the model is shown in Figure 17. 

Figure 17, Acrylic Model of Oil Passages 

Figure18, Main Oil Galley 

Oil distribution in the new cylinder block is from a 5/16 inch diameter full-length horizontal main oil galley in the valve chamber located between the cylinders and tappet bosses. Figure 18 (a sectional view taken at the centerline of the front main bearing and looking towards the rear) shows the location of the boss for the main oil galley. It also shows a diagonal groove running from the main oil galley boss to the front main bearing that will become a boss and be drilled to supply oil. Similar bosses are also designed into all five main bearing webs to accommodate drilled oil passages that will supply pressurized oil from the main oil galley to both crankshaft and camshaft bearings. Main bearing oil passages will be 1/4 inch diameter and camshaft oil passages will be 1/8 inch diameter. Model A and B engines have pressed in steel tubes for oil passages that are not on the centerline of the main bearings. This new design cylinder block has the oil passages centered on the main bearings.

 Other bosses that will become oil galleys to route oil from the oil pump to the main oil galley are provided. These bosses (oil galleys) are designed to permit oil routing with or without an external oil filter and they have been positioned to be in line with specific bolt hole locations for ease of plumbing. For instance, if an external filter is desired, oil exits through the stock pipe plug found on the oil pump side of the cylinder block, goes through the filter and then is routed back to the main oil galley through a series of drilled passages that are inboard of the lower bolt that attaches the timing gear inspection cover (A-6017). Of course, this lower bolt (A-21111) would have to be drilled and tapped for the return oil line from the filter. On engines not using an oil filter, all routing is internal through drilled passages and nothing shows from the outside. Whatever the routing, there are no pipes, tubes, hoses or fittings internal to the new cylinder block.

 The vertical height of the new main oil galley was positioned at the same height as the lower row of valve chamber cover bolts. Bosses are provided and drilled behind two of these lower cover bolts. The purpose of this is to allow a pressure tap for an oil pressure gauge that will read directly from the main oil galley (If the lower cover bolt is drilled and tapped for a fitting).

 On a pressurized (modified) Model A engine, the oil pump must be sealed to the cylinder block and fastened in place. Sealing on the new cylinder block is accomplished with an o-ring that fits in a groove in the cylinder block bore where the oil pump plugs into, and a boss is provided in the center main web for a bolt and dog type clamp to fasten the oil pump.

 The front and center main bolts also have an o-ring seal in the cylinder block to keep pressurized oil from leaking up and out at their castellated retaining nuts. Extra material has been added to the main bolt bosses to accommodate the o-ring grooves.

 Figure 19 is a view of the underside of the valve chamber cavity core and shows the recess (between cylinders and tappet bosses) that will become the full-length oil galley boss. 

Figure 19, Bottom View of Valve Chamber Core 

ASSEMBLY OF CORES

 Up until now, the assembly of cores has been shown in an upright orientation for simplification. In reality, the cores would be assembled upside down. This is for easier placement of valve port cores into their core prints. Once positioned, all cores would be held in place with cement. Twenty-one cores (seventeen different) are required to create the mold for the cylinder block. 

Figure 20, showing the assembly of top, upper water jacket, and valve port cores. 

Figure 21, showing the right side and lower water jacket cores added. 

Figure 22, showing the valve chamber, distributor shaft, and cylinder 4 cores added. 

Figure 23, showing the cores added for cylinders 3,2,1, timing gear cavity, and the front. 

Figure 24, Completed Cylinder Block Mold 

Figure 24 shows the completed mold with the left side and bottom cores added. Note that the bottom core prints go completely through the bottom. This is to allow alignment of cores as the bottom is placed in position.

 To make a casting, molten cast iron would flow through a system of sprues, runners, and gates to fill the void areas in Figure 24. Gates into the casting through which molten iron flows will be placed at locations that will later be machined. After solidification, the cores which are made of green sand (foundry term) are broken out, and the gating system is cut off. What remains is the raw unfinished casting ready to be machined.

MACHINED CASTING SOLID MODEL

 A solid model has been created that accurately locates all machined surfaces and holes relative to each other within 8 decimal point accuracy. This model was created independently from the crankcase cores so that when superimposed on the crankcase cores, it will serve as a check to verify the validity of the design. This solid model uses six different colors to indicate which features are machined from each of the six sides of the casting. This solid model’s only deviation from the drawing (A-6015) is that the exhaust ports are not counterbored for the gland rings (A-9440). Figures 25 and 26 show two different views of this model. This model is machine shop friendly and very easy to transform into a CNC program. 

Figure 25, Machined Casting Solid Model 

Figure 26, Machined Casting Solid Model 

Figure 27, Section at Center Main Looking Towards Rear 

When the machined surfaces model is superimposed into the completed mold, various section cuts can be made to insure that there is adequate machine allowance and that all drilled passages and threaded holes are centered in their bosses. One such section is shown in Figure 27 which is a section taken at the center main looking towards the rear. 

OIL PUMP DRIVE BEARING DESIGN (A-6560)

The oil pump drive bearing (A-6560) has been redesigned to make it compatible with the requirements of an engine with full oil  pressure. The oil pump drive gear (A-6551) is unchanged. These changes to A-6560 include an o-ring groove and a radial lip seal (National #330699).  The new oil pump drive bearing will be held in place with a setscrew that is hidden behind the valve chamber cover. 

OTHER PARTS

 Oil pump pressure should be regulated to keep loads between camshaft and oil pump drive gears reasonable. Several aftermarket oil pumps have regulators built in. A stock oil pump can be regulated if the bottom plate is replaced with a new bottom plate housing containing the regulator.

 Flywheel inertia should be reduced. A 22-pound flywheel featuring a cast iron center section and friction surface with an aluminum bell to support the ring gear is in production. This flywheel utilizes the 48/09A-7563 pressure plate used in the V8 Ford from 1935 until 1942. Figure 28 is a photograph of this flywheel.


Figure 28, 22-Pound Flywheel 

COMMENT ON MACHINE SHOPS AND REBUILT ENGINES

Automotive machine shops are different from conventional machine shops. 

The machinery found in a conventional machine shop can take a piece of metal and make a part from it. Feature location, dimensional accuracy and surface finish are all important. 

The machinery found in automotive machine shops is very specialized and used to create new wearing surfaces on worn parts. In general terms, automotive machinery cannot make the original part. Dimensional accuracy and surface finish are important, however feature location is not as important or controlled. Automotive machine shops also provide a variety of repair services such as cleaning, detecting and repair of cracks, installing thread inserts and welding.

 In a conventional machine shop, the accuracy of the part being machined is directly related to the accuracy of the machinery doing the cutting, and the quality of the part can be determined by measuring it and comparing it to the drawing from which it was made.

 In an automotive machine shop, previously machined surfaces are usually used for alignment. Accuracy is dependent on both the machinery and the location of previously machined surfaces, and can diminish with each succeeding rebuild. An automotive machinist is forced to make assumptions that previous rebuilds and machined surfaces are accurate. Unfortunately that is not the case for many engines. 

A new Model A engine had very closely controlled feature locations, dimensions and surface finish. Over the years, and after many rebuilds (especially with modified and "improved" engines), the features are not necessarily in their correct locations. 

During the Great Depression when Model A’s were just old used cars, many engine rebuilds were done by "shade tree" mechanics with substandard equipment. Ford factory rebuilds were held to high standards and dealer rebuilds using K. R. Wilson equipment were second best.

 Take cylinder bore as an example. With a fresh casting, the manufacturer (Ford, or a conventional machine shop) will first machine the bottom of the cylinder block. Next, this bottom surface is clamped to a machine table and the cylinder is machined. The cylinder is perpendicular to the bottom of the cylinder block, which is clamped to the machine tool table. Please note that the bottom of the cylinder block also happens to be the crankshaft axis centerline, which means the cylinder will be perpendicular to the crankshaft. Accuracy is dependant on the machine tool.

 After years of use and when a worn cylinder bore needs to be rebored, an automotive machine shop is called upon. To rebore a worn cylinder, the typical automotive machine shop will clamp a boring bar (the machine that rebores the worn cylinder) to the upper surface of the cylinder block, and then determine the position of the new cylinder bore by locating the new oversize cylinder bore from the unworn bottom portion of the worn cylinder bore. Next, the cylinder is bored and is perpendicular to the upper surface. The cylinder diameter is exact and surface finish looks great.

 However, two important dimensions are not controlled. These are perpendicularity to crankshaft axis and cylinder location. What if in 1934, Uncle Clyde fixed a crack in the top of the cylinder block, resurfaced the upper surface of the cylinder block with a belt sander, and then rebored the cylinder perpendicular to this new surface that he created? How would a present day automotive machinist know this, and how could he compensate? Even if the upper deck is resurfaced, is the cylinder in the right place? Accuracy is dependant on a number of unknown factors.

 Similar stories can be told regarding crankshaft axis miss-location in the crankcase, miss-location of connecting rod journals relative to the woodruff key at the front of crankshaft that determines valve timing, and flywheel flange runout relative to rear main journal diameter.

 Balancing is making things equal. New engines have all of their moving parts individually balanced and there is very little crankshaft flywheel mounting flange runout so that any order of assembly will result in a balanced assembly. Rebuilt engines with crankshaft flywheel mounting flange runout must have the crankshaft and flywheel balanced as an assembly and match marked for proper reassembly. Switching flywheels from one rebuilt engine to another will result in an unbalanced assembly. 

Since the accuracy and exactly what was done during previous rebuilds is unknown, it should be no surprise that two different rebuilt engines from the same automotive machine shop held to the same dimensional tolerances and using the same machinery can have two entirely different personalities. There is some truth to Jonathan Swift's adage, "You can't make a silk purse out of a sow's ear" 

ASSEMBLING THE NEW ENGINE 

One advantage to assembling an engine from the new parts described in this article is that they would be ready for assembly and require no machine work. Building a new engine would simply consist of ordering a set of new parts, cleanup of attached parts (manifolds, head, covers, oil pan, oil filler, cam, timing gears, and valve train), and assembly.

Upon assembly, up to three setscrew type plugs would have to be removed from the ends of oil galleys in the new cylinder block depending on whether a three or five journal camshaft is used and whether an oil filter is used.

 Due to the fact that this engineering study and new design was constrained to maintain all interfaces and basic Model A engine dimensions, other assemblies are possible. 

For instance, if someone wanted to just implement the new crankshaft and connecting rods into an original Model A cylinder block, it could be done by machining the three main bearings to a smaller diameter, removing the extra thickness of the flywheel mounting flange, and adding dippers to the connecting rods. The new crankshaft would not have the support of the added intermediate main bearings, but it would still be stiffer than a stock crankshaft and would have engineered counterweights.  

Or, if someone wanted to build an all out race engine, they could take the new cylinder block (5 mains, fully pressurized, and super rear main seal) and add their own crankshaft, connecting rods, and wrist pins made from titanium alloys, crankshaft counterweights made from platinum (platinum is denser than tungsten), and pistons made from a magnesium or beryllium alloy. 

 FILLING A VOID

Today and in the future, original Model A engines will continue to be rebuilt, and many will continue to be modified. This new design is intended to fill a void in the Model A hobby. If it goes into production, builders of hot-rodded engines will have a better foundation from which to start, and restorers wanting to drive their car hard can preserve their original engine by removing it and storing it for later use. If this engine goes into production, skilled restorers will be able to assemble their own engine and rebuilders will be able to offer their customers a product with a guarantee without worry from comebacks. The price of good used original cylinder blocks and crankshafts will hopefully drop because new ones are available. And hopefully, parts that have been hoarded for years will become available for sale at reasonable prices. 

STATUS OF ENGINEERING


This engineering study is complete. It proves that an engine can be built to modern standards by changing a minimum number of parts. Preliminary calculations indicate that that the stresses at 150 HP and 5000 RPM are within the allowable limits (fatigue endurance limit) for commonly used automotive materials (Ductile iron connecting rod and crankshaft, gray iron cylinder block). Stronger materials (added cost) would easily allow higher HP and RPM without any redesign. 

Additional processing (added cost) by rolling the fillets and nitriding would make the ductile iron crankshaft stronger (however this added strength would be lost if the crankshaft were ever reground in a later rebuild). Ductile iron connecting rods could be made stronger by shot peening.

The creative engineering for this study is complete, but the detail engineering is not complete. 

The detail engineering tasks that still need to be completed include additional FEA analyses, small design changes as a result of the FEA, creation of drawings with notes and dimensional tolerances, selection of materials, specification of processes, and surface finishes. Engineering effort is also required for meetings with foundries and machine shops, and to write a Quality Assurance (QA) plan to assure that any hardware delivered meets all solid model, drawing, material, and processing requirements.


 The cylinder block casting cores are all designed and documented with solid models. What remains is a small amount of engineering at the foundry to add draft and shrinkage allowance, make core boxes, and design the gating system. Engineering is also required at the machine shop to transform the "machined casting solid model" mentioned above into a program for CNC machines.

 Several small cylinder block external discrepancies between the original Ford drawing dated 28 March 1929 and several original cylinder blocks produced were noted (mostly differences in fillet radius and centering of oil pan attachment bolt holes in their bosses). Engineers in the Model A era were mainly concerned with machined interfaces and what happened between these interfaces was often left to the discretion of pattern makers. The solid models created in this study are from the original drawing, but modified in an attempt to reflect actual hardware produced, but they are not 100% correct. With today’s technology, the external surfaces of an original cylinder block can be laser scanned to exactly create a solid model that is 100% identical to an original cylinder block.  Cylinder block *A4619726* is an extremely good example because all of its’ oil pan attachment bolt holes are centered in their respective bosses. So, for this new cylinder block to match an excellent original example, the exterior of cylinder block #*A4619726* should be laser scanned and this laser scan will become the solid model of the exterior of the new cylinder block.  

The redesign presented and described in this article is not optimal. If time and cost were no object, the design / FEA cycle could go through many iterations to finally optimize the design. Intake and exhaust port design could be improved upon with flow bench testing. Additional optimization could easily be achieved even with today’s materials. Exotic processing methods and coatings could be used to enhance performance. Metal Matrix Composites (MMC’s) and ceramics could also be utilized. 

WHAT’S NEXT

This study and engineering effort is on temporary hold. It can end now or it can go on to produce hardware. 

This new engine is not something that would be put into continuous production, but would instead be built in batches (production runs). Foundries and machine shops must work in quantities to keep costs reasonable. Setup and tooling costs are similar whether 1 or 10000 parts are produced. For a viable bid, foundries and machine shops need to have a quantity number.

Target cost is under $3000, and hopefully less for this new cylinder block assembly that also includes a crankshaft with connecting rods, seals, o-rings, bearing inserts, and fasteners. New replacement small block Chevy short blocks from China with ductile iron crankshafts and connecting rods typically sell in the $2000 range. Actual cost would depend on quantity ordered and whether the parts are manufactured in the USA or China. This target cost compares favorably with the cost of a rebuild, is a fraction of the cost that many have spent on modifications and improvements, and should be within an acceptable price range for someone collecting parts for a future engine. 

Factories all over the world are ISO (International Standards Organization) certified and theoretically they are all equivalent, which means that anything produced by any ISO factory is equivalent. In China, tooling is typically disposed of after every production run, and any subsequent production run would have to pay for all new tooling. Tooling in America is a little different. The customer who paid for it usually owns tooling in America, however it is often not compatible from one producer to another.

The first production run of this design (and any succeeding runs) should be at least 500 units to spread the costs of tooling and setup to keep costs reasonable. Larger quantities would result in lower pricing.