We do nothing at all in what is sometimes ambitiously called research, excepting as it relates to our single objective. We believe that anything else would be outside our province and possibly done at the expense of our own particular function which, to repeat, is making motors and putting them on wheels. In the engineering laboratory at Dearborn we are now equipped to do almost anything that we care to do in the way of experiment, but our method is essentially the Edison method of trial and error.
As it is, our task is rather a large one, for we must look well ahead to the possible depletion of sources, to the saving of material, and to the finding of substitute materials and fuels. Quite often, we merely put the results of our experiments away for future use in case market conditions should change. For instance, if gasoline should go above a certain price, then it would be practical to bring in substitute fuels. But our principal duty, as we conceive it, is not to wander from our own path, but to learn to do one thing well. Learning to do that one thing well has taken us into many fields. We want to save material and we want to save labour and scarcely a week passes in which some change is not make. Some are of minor and others of major importance, but the methods of procedure is always the same. Curiously enough, some of our largest savings come in the manufacture of parts where we thought we were doing rather well.
In one case we found that by using two cents more worth of material in a certain small part we were able to reduce the total cost of it by 40 percent. That is, the amount of material under the new method cost about two cents per part more than under the old, but the labour was so much faster that, under the new method, the cost which was formerly $.2852 was not only $.1663 — we carry our cost out to four decimals. The new method required ten additional machines, but the saving was nearly twelve cents per part — that is, the cost was almost cut in two which, on a 10,000 a day production, meant a saving of $1,200 a day.
From the beginning of manufacturing until several years ago we had used wood for the steering wheels. This seemed a great waste, for only the best quality of wood could be used and no wood-working operation can be carried through with absolute precision. At the same time, out on the farm at Dearborn, we had tons of straw yearly going to waste or being sold for next to nothing. Out of this straw we developed a substance which we call Fordite, which looks like hard rubber but is not. The steering wheel rim, and, in all, about forty-five parts of the car, mostly having to do with the electrical work, are now make out of this straw, and the production is so large that the farm will produce only enough for about nine months. Then we have to buy straw. This is the process:
The straw, rubber base, sulphur, silica, and other ingredients are mixed in batches of 150 pounds each, which then go to the rubber mills, where they are mixed in heated rollers for forty-five minutes. Then the mass is fed into tubing machines in small strips and comes out through a round die, much as sausage from a grinding machine. As it comes out it is cut, on the bias, in lengths of fifty-two inches and then is ready to be rolled into an outside covering of fine rubber-like substance. This is then put into a mould under hydraulic pressure of 2,000 pounds to the square inch and heated by steam for nearly an hour. When they come out of the heat, the wheels are soft, but they soon take on a flit-like hardness that remains.
Next, these steering wheels go to the finishing rooms, where they are smoothly trimmed and polished. The pressed steel “spider,” or cross piece, is then placed in the wheel and securely fastened by a machine which in one operation bores a small hole and in the next screws in the screw. The steering wheel is then ready for shipment and final assembly on the car.
We save about half the cost of wood — and we conserve wood.
The touring car uses about fifteen yards of artificial leather for the top, curtains, and upholstery, and we need altogether five grades. Using natural leather would be quite out of the question. In the first place, it would be too expensive, while, in the second place, not enough animals are slaughtered to begin to provide for our requirements. Our people had a hard time developing an entirely satisfactory artificial leather — it took them five or six years. First, they had to get the proper coating compound for the cloth which is the base of the leather, and then to make the operation continuous. Making our own leather not only renders us independent — which was the original purpose of the undertaking — but also saves us more than twelve thousand dollars a day. Essentially, these are the operations as we now perform them:
The cloth is fed into ovens. The ovens consist of a series of towers. At the base of each is a tank containing the coating compound. This is poured on the cloth as it travels through, a knife spreading it evenly and scraping off the surplus. After receiving the coating, the cloth ascends the tower to a height of thirty feet, at a temperature of about two hundred degrees. By the time it has descended, it is thoroughly dry. The second oven gives it another coat, dries it in the tower, and brings it down to the tank in No. 3 oven, and so on, until the first seven coats have been given.
It is then weighed to determine the amount of coating per running yard and sent to the embossing press, where it receives the graining under a pressure of 700 tons. One more oven gives it the finishing or sealing coat, adds luster, and keeps the material pliable.
The compound is a mixture of castor oil and drop black mixed with a preparation of nitrated cotton dissolved in ethyl acetate and thinned with benzol. This is highly volatile, which accounts for the easy drying. The fumes of the ethyl acetate, alcohol, and benzol are driven off in the ovens, but — they are recovered by a special apparatus we developed. The fumes are drawn through charcoal made from cocoanut shells until the charcoal becomes saturated. Steam is then turned in, which drives the fumes into a condenser, from which they are separated into the original compounds. As much as 90 percent of the fumes have thus been recovered when the work of the condenser has been concentrated on one smokestack. The manufacturing is continuous. As soon as a roll of cloth is nearly used up, the end is unrolled by hand and sewn to a new bolt. Thus the coating continues without interruption — an important factor when one considers that even a brief delay would cause the compound to harden on the knives.
There are no lights within the building, all artificial illumination being furnished from the outside on account of the fire hazard. Every machine is grounded, and as many precautions against fire taken as in an explosive factory, and we have had no accidents at all.
The treating of steel by heat is of the highest possible importance, for it makes possible the use of lighter parts by increasing their strength. But it is a delicate process: a part must not be too soft or it will wear out, or again, if it is too hard, it will break. The exact state of hardness depends upon the use to which the part is to be put. This is elementary. But the treating of large quantities of parts so that each will be of the right hardness is far from elementary.
The old way was to guess. We cannot afford to guess. We cannot afford to leave any process to human judgment. In our former heat treat processes, we thought that we were fairly advanced. And we were advanced for the time, because the work could be done by men after only a little training and the results were uniform, owing to the mechanical regulation. But the heat treat departments involved hot, hard labour, and we do not like to have jobs of that sort in the shops. Hard labour is for machines, not for men. And also the straight parts, such as axles, did not cool evenly, and after treatment they had to be straightened, which added to the cost.
We set a young man the task of bettering all our heat treat operations. He felt his way for a year or two and then began to get results. He not only cut down the number of men, but he devised a centrifugal hardening machine which cools the shafts evenly all around. Thus, they do not bend, and the straightening operation is no more. The electric furnace replacing the gas furnace has been one of the large steps forward. Where four gas-fired furnaces, with six men and a foreman, did 1,000 connecting rods an hour for the drawing operation alone, now two electric furnaces will both harden and draw 1,300 rods an hour, with only two men — one to feed and the other to take off.
For the heat treat, the axle shaft department uses a large two-deck furnace. A walking-beam, working slowly, moves the shafts forward into the lower chamber of the furnace at intervals of one minute. It takes twenty-eight minutes for a shaft to move completely through the lower chamber of the furnace, and during those twenty-eight minutes it is in a constant heat of 1,480 degrees Fahrenheit, the temperature being regulated by instrument control.
As the shafts slowly come out at the far end of the furnace they are seized by an employee with tongs and placed one by one in a spinning machine. They are quenched in caustic solution at the rate of four per minute; the spinning motion given them by the machine makes the decrease in temperature practically instantaneous over the shaft’s entire surface. This operation goes to insure a uniform hardness and avoids pulling them out of shape by uneven cooling.
The quenched shafts are carried by a conveyor to the upper chamber of the furnace, and move back toward the entrance end through a constant heat of 680 degrees Fahrenheit. It takes forty-five minutes for this treatment. Thoroughly drawn, they are sent by overhead conveyor to the final machining.
These changes may not seem important, but cutting out the item of straightening after the heat treat has saved us around thirty-six million dollars in four years!
We investigated the making of electric storage batteries and, after a period of trial — we always try out everything thoroughly before we go into it — we found that we could make batteries cheaper than we could buy them.
For the car and the truck 162 steel forgings are needed, and this has developed a forging department which daily uses more than a million pounds of steel, and in which, by constant change and experiment, we have saved many millions of dollars by combining in one operation complex forging operations which formerly required several, and also by extending the use of upsetting machines — that is, machines which press the steel into shape instead of hammering it. Our objective is always to minimize the subsequent machining.
In one of these upsetting machines, heavy dies placed in vertical order crash heavily together upon the heated steel bar. Three operation sets or more are needed, except in a few instances, thoroughly to shape a bar of steel to its required form. The bar is inserted between the top dies first. These do the actual upsetting — that is, strictly, the thickening and shortening of a portion of the steel bar to the degree required. In rare cases, two sets of upsetting dies are necessary. The remaining sets of dies shape, pierce — if necessary — trim, and cut off the shaped portion. The steam hammer group has ninety-six hammers. The smallest hammer in the group carries a ram and piston of 800 pounds weight, and the largest strikes a blow with a ram and piston weighing 5,600 pounds.
Dies are set in the anvils and hammer-faces. As in the case of the upsetting machines, each hammer is set with dies that enable it to perform a complete phase of the work of manufacture. There is no division of labour between hammers. In the forging of a crankshaft a bar of hot steel is placed across a bending die at the left of the anvil; a blow from the corresponding die on the hammer-face wrinkles the bar into the semblance of a crank shaft. The crank-bent bar is moved to the right; several blows from a second die complete the resemblance — the result is a crank shaft, in the rough, though the two ends are of equal size, and the whole forging is framed in a thin flashing of excess metal. The hammer work is complete. The flashing is removed on a trimming press, and the flange at the crank shaft’s end will be formed on a upsetting machine.
Some parts require only a portion of the bar for their manufacture. On the hammers where these are made, a small cutting block is placed, and a blow from the hammer separates the shaped portion from the remainder of the bar. Smaller forgings are made on hammers whose dies contain several exactly similar impressions, permitting a number to be forged at once.
The order in which forgings requiring both hammer and upset work are manufactured varies according to the peculiarities of the part to be formed. The axles go first to the upsetting machine, which shapes them roughly, spreads and divides their ends; from there they go to the hammers. Half their lengths are formed, made ready for machining, at a time, since they are too long to be put under the hammer entire.
To rid the forgings of the flashings that border them when they leave the hammers, eighty trimming presses are used. Most of these presses straddle a belt conveyor, so that the flashings that are trimmed from the forgings are carried away immediately. Small forgings are also allowed to fall to the conveyor. Near where the conveyor makes its exit from the building, these forgings are removed and sorted into boxes. The conveyor discharges the flashings into a car on the switch outside.
In the trimming presses, the punch contains the form of the forging in relief, while the die contains an opening exactly corresponding to it. The forging is thrust by the pressure through the opening, while the flashing is left lying on the surface of the die.
Special devices are used to make and keep the various longer forgings accurate in length. A separate mechanism is used for this purpose with the axles. The steering post, during its formation on an upsetting machine, is held to an accuracy never deviated from more than a thirty-second of an inch.
Thirteen upsetting machines are now used on the triple gear job. In making this detail it was once necessary to shape three separate forgings. It is now forged from a single billet of steel.
The most difficult upsetting job in the drop forge is the drive shaft roller bearing housing. This calls for a double upsetting and rather elaborate shaping at both upset ends. Nevertheless, it is being successfully accomplished on a single machine.
An interesting piece of equipment is the reclaiming steel rolling mill. In this mill, remnants of stock too short to be used are reduced in diameter and increased to usable length by successive trips between progressively smaller rollers. This salvaging is done on the spot to save transport.
The casting of aluminum in dies has made for considerable economy. It took us some years to devise a satisfactory method. For a long time die casting was looked upon as impossible. The old method of casting in sand moulds lets the air exude through the sand as the hot metal is poured in, while pouring metal into dies or solid moulds caused air bubbles to form, resulting in “blows” in the cast. Then the secret of feeding the molten metal into the dies from underneath was discovered.
The die is placed directly above the pot containing the molten metal. In fact, it takes the place of a lid. When a cast is to be made, the operator turns the air pressure into the hot metal. The pressure forces the metal up through a feeder into the die. As the metal goes in, the air is forced out through minute vents. As the top of the die is filled first, the cast naturally solidifies from that point downward. The air is forced out by the first rush of molten metal on the top of the mould, and as metal can enter only by the feeder, all danger of air bubbles is eliminated, and the cast is perfect.
Insulated copper wire — and we use large quantities — is expensive. So we set about making our own and now produce about one hundred miles a day. We use standard wire-making machinery, but with many improvements and simplifications made possible by using the machines for a single product. The process starts with 5/16 inch copper rod stock similar to that used for trolley wires. This is drawn through nine successively smaller chilled iron dies. Wire from the last die is about 3/32 inch diameter, traveling to the winding spool at a speed of 725 feet per minute.
The drawing process causes much heat, which is carried away by water flowing over the dies. It also tends to make the wire hard. To soften for further drawing, the wire is plunged into water on a turntable which is revolving until the load is under the furnace chamber. It is then raised into an airtight cylinder to be held at a temperature of 1,045 degrees Fahrenheit for one hour. Air is excluded to prevent oxidation.
Machines for the second drawing are fitted with eight pierced diamonds through which the wire is drawn, each reducing the size a few thousandths of an inch. These diamonds, which may cost $300 each, can be used six months without appreciable wear. The final die, .044, produces twelve-gauge bare wire ready for insulating.
The insulation consists of five coats of dielectric enamel and a wound cotton covering. Enamelling is continuous and automatic. Four men easily take care of eighty rolls of wire at once, as they are unwound and rewound with each enamel coat baked on at 845 degrees Fahrenheit. The enameled wire, every inch inspected for roughness or breaks in enamel, is passed to the winding machines. Bad stretches of wire are cut out and the ends brazed and reenamelled.
Cotton-winding machines have prepared eighteen ply or “end” cotton wound on bobbins for the insulating machines. These bobbins whirl around the wire as it passes through them, winding an even coating of cotton at great tension over the enameled surface. Four men here take care of the work of seventy-two spindles. The machines are almost entirely automatic.
The screw driver is an ancient and valuable instrument, but one man with one screw driver is hardly in line with modern methods. He can scarcely do enough work to pay his way. We are trying to get away from the screw driver. For instance, we now have a sixteen spindle screw driver, which drives home sixteen screws into the starter ring gear in a single operation.
As the transmission comes along, the conveyor bolts and washers are assembled through the sixteen magnet heels, and the bolts turned about, one thread into the flywheel holding the magnets in position. A white metal spool is placed under each magnet end, and a magnet clamp set on top. A brass screw is inserted through a hole in the magnet clamp, passing between the magnet ends, through the white metal spool and a small hole in the flywheel, and into the starter ring gear. All is now in position for tightening the screws and bolts.
The transmission slides under the spindle screw driver and, with a slight movement of the operating lever, the locating arm is dropped. The locating arm has a notched edge which fits over the four transmission bolts, bringing the screws on the rim of the flywheel directly under the screw drivers, which are suspended from a circular spindle guide plate held in position by a movable head. Each screw driver is encased in a thimble which drops over the screw head and guides the screw driver into the slot, as the motion of the lever is completed. As the screw is driven in, the friction increases, more power being used for the last turn of the screw than for the first. When the screw has been driven home, a friction clutch, encased in the spindle arm, slips, action ceases, and the screw slots are prevented from being broken off.
From the spindle screw driver, the transmission passes to an eight spindle bolt driver, which works on the same principle as the screw driver. The bolt driver tightens the bolts which pass through the heel of the magnet into the flywheel. Before the sixteen spindle screw driver was placed in operation, six men were required to tighten the screws. Now the work requires but one man, and the operation is completed in a few seconds.
Following the same thought is the use of rivets instead of screws in putting together the body parts. The rivets give better service than screws, and they can be put in faster — they will go in still faster when we have developed the magazine riveter which we are working on. We use 3,000,000 rivets a day.
The methods of casting bronze bushings have been constantly changed and bettered in order to get rid of hard hand labour, until now the department has almost nothing about it to suggest a foundry. The melting process is carried on chiefly by twelve electric furnaces, each taking a ton of metals at a charge and requiring about seventy minutes to the heat. The furnace is left motionless until the metals are thoroughly melted, then a gentle rocking to and fro gives a uniform mixture. When the molten metal reaches a temperature of 2,200 degrees Fahrenheit, a sample is sent to the testing laboratories to be analyzed, while the remainder is emptied into small buckets lined with fire clay which are trundled by an overhead supporting tram to the pouring line.
The patterns used for casting small bushings resemble several bushings united by gates with four or five gates united at one end to form a good-sized cluster. By using such a pattern, only one mould and one pouring operation are necessary to produce a larger number of bushings.
The moulders are provided with all possible aids to rapid and efficient work. Instead of sifting the same by hand over the pattern, electric riddles are used — a press on a button does the work. The sand must be shaken and packed down to make a solid, dependable mould. Here, again, machines do a far better job than any workman could do. Electric coils under the metal table heat the mould that is being made without producing unnecessary warmth for the worker on a hot day. In line with this idea is the system of cold air blowers which furnish streams of cool air near each workman. Finally, in order to remove the pattern, the mould must be in two parts fitting closely together and yet easily separable. Between the two halves, it was once customary to spread lycopodium, a fine powder made from the pollen of flowers found only in Russia. This was very expensive, but now we have a cheaper preparation which is just as effective. An air vibrator and a simple gear arrangement make it possible to lift the upper half of the mould evenly, so that it will not be damaged in the process.
The finished mould is carried by a continuous conveyor to the pouring line and filled with the liquid metal. There is a tendency at this time for the mould to separate, allowing the metal to seep in between the halves. Formerly it was necessary to place heavy weights on the moulds, which was an occupation for husky workers. Now, however, a simple clamp, revolving on the spindle supporting the platform on which the mould is riding, does the same work with a single stroke of the hand. Farther along, the moulds are broken and the hardened castings removed, while the steel flasks used to support the outside of the moulds are returned by the same conveyor to the moulders.
The clusters of bushings are broken apart and the connecting gates sent to the charging room. The bushings are then placed in large cylindrical mills and tumbled about until the sand is cleaned off, and finally they are placed in the notched rim of a wheel which carries them against a grinding disc to smooth off the knob left by the gates. The castings are now ready for the finishing rooms, but are not sent through until the fracture test results come from the laboratory. The castings made from each heat of metal are kept separate from one another until this test has been made. Transportation costs and delays are lessened by locating the three finishing departments directly above the foundry. Practically all of the machinery for finishing is automatic and foolproof. Double automatic lathes turn out six thousand piston pin bushings every eight hours with such accuracy that only 1.3 percent fail to pass the inspection requirements; automatic punch broachers finish the bore directly from the rough castings for these piston pin bushings as well as the larger ones: automatic drills handle a few types that cannot be easily broached; even the inspection is automatic.
The automatic lathe needs but one speed and one size of arbor. Instead of being thrown out of gear every few seconds, the direction of the feed is reversed, and rather than have the machine run idle while the finished bushing is being replaced by one in the rough, two arbors are supplied, the cutting tool operating between them like a shuttle.
The broaching machines have their most dangerous feature — that of mistaking the workman’s hand for a piece of metal — eliminated. Instead of its being necessary for the operator to reach under the punch, a metal channel is provided which is kept filled with castings. As the broach rises, a push on the end of the line moves the first one into place. The broached bushing drops into a chute while the next one is fed in through the channel in a fraction of the time required by the hand method of feeding. The automatic drill presses are in duplicates like the lathes, but in this case there are two drills, one rising while the other is cutting.
For sorting or inspecting as to length of the bushings, there is a machine with three sets of discs, so arranged that the first pair will take up bushings which are too long, the second pair will remove only those which are within the limits of the specified length, while the third pair will take the undersized. This is all regulated by the distance between the discs, and this may be adjusted by ten thousandths of an inch. The bushings slide into notches in the rim of a large wheel that holds the pieces in line for the sorting discs.
The outside diameter sorter is even simpler. Two ground and polished rollers are set parallel to each other and on an incline, the diameters of the rollers decreasing by steps toward the lower ends, so that the space between them increases. Bushings are fed on to the rollers by gravity and are revolved. As they travel down the incline, they fall through into the space underneath which is so divided that the undersized ones go into one chute and the good ones into another. Those which remain on top are dropped from the end of the rollers into the over-sized chute.
What does all this mean? This: in 1918 this entire department averaged 350 finished pieces per man per day of eight hours, with a machine scrap of about 3.0 percent. At present 830 completed pieces is the daily average, while only about 1.3 percent of the product is scrapped.
In the making of springs, a similar advance has come, both in accuracy and man saving. In shaping the leaves, the forms used keep them so exact that they are interchangeable with corresponding leaves on other springs. The leaves are formed and hardened in oil in one operation. Next, they are tempered in nitrate at 875 degrees Fahrenheit, after which they are graphitized and used.
In 1915, the department employed four men to make fifty springs a day; at present, 600 men make 18,000 springs a day.
We must have inspectors at every stage of the work; otherwise, faulty parts might get into the assembly. Our inspectors in only a few cases are required to use judgment — mostly they apply a gauge, but, as was shown with the bushings, we are working toward mechanical inspection. For instance, electricity at 20,000 volts now tests the timing of the eight cams on the Model T Ford camshaft, not only more precisely than was possible by the former method, but seven times faster. Operated by one man, the new electrical gauge displaces seven of the old type gauges and their operators. The electrical test takes ten seconds.
In the new gauge, the camshaft is inserted in bearings, so that the cams operate push rods just as they do in the assembled motor. Instead of operating against valves, however, the push rods in the gauge close and open electrical contacts as the shaft is revolved. These electrical circuits are supplied with current only when the opening and closing points on the contour of each cam are in contact with the push rods, a distributor in the handwheel by which the shaft is revolved taking care of this.
If, at a critical position of the cam, the contour is too high or low within very close limits, and electrical contact is made completing a circuit which causes an electrical indicator to flash. There are two of these indicators, one for high and one for low cams. On the handwheel is an index, the position of which at the time of the flash indicates which cam is faulty. If the cams are all accurate within specifications, the electrical indicator does not flash at any point during the revolution. The electric gauge may be set to detect errors of two ten thousandths of an inch.
But this is the sort of thing which is going on every day — we take it as our duty to use the public’s money to the advantage of the public by pressing always for a better and cheaper product.