Ed: The terms in this little essay are made-up or reverse-engineered, so they aren’t the official terms used in machine design. I’m not an engineer. This is the introduction that I would like to have read years ago but have been unable to find despite searching. A good book with better information likely exists somewhere out there among the billions of books, and I’d love to hear about it if you’ve found it.
The main reason we’re seeing a loss of proficiency in engineering and the trades is the dominance of anti-logical attitudes toward learning about them. The prevailing attitude is that one either “just gets” how machines work or “just doesn’t get” it. I.e. There are natural pick-up artists, and everyone else should resign themselves to involuntary celibacy. A similar attitude can be seen in music, where people assume that real musicians just pick up instruments and figure them out naturally.
This has been the common perception for math for at leat my entire life
It comes from an antipathy toward people who show EFFORT. Low-effort humans discourage effort in others and promote narratives of low-effort successes and tear down high-effort practice regimens. When they see obsessive practice, they ascribe it to a sort of possession. This, they understand, would be a form of low effort since it entails “simply letting go” and letting the possessing spirit have its way with you. This is mostly nonsense.
Understanding machines has more in common with syllogisms and computer programming than it does with artistic composition. It is actually about as easy as syllogisms and a great deal easier than computer programming. But you have to have a solid understanding of the fundamental concepts, which as far as I can tell are no longer taught (since it’s assumed these are understood by instinct, or not at all). This has led to verbally inclined people fleeing from what was once thought of (by the Victorians) as a profession for natural philosophers and intellectuals. Enough ranting, on to the fundamentals.
1) Machines are designed beginning with the counterfactual and working backwards.
E.g. “I want to apply 5,000 pounds of force to punch a 1/4-inch hole in a 1/2-inch steel plate.” Everything else in a machine is just enabling the final action of punching the hole. That could be motors, power supplies, compressed air filters, all kinds of things that only exist to enable the machine to do the one, particular thing it exists to do.
2) Machine operations are designed as step-by-step verbal descriptions first.
-The operator pushes a button.
-The button has a piece of wire under it that, when pushed connects to pieces of wire on either side.
-That completes a circuit which connects a battery, a light bulb, and the button in series.
-Current is then allowed to flow through the button to the lightbulb and back to the battery.
-As the current enters the lightbulb it experiences resistance trying to pass through the filament. The filament heats up.
-The heated filament glows, producing light.
You can continue this process of describing what causes what forever if you want. Usually the description only goes deep enough for the practical purpose of the designer clearly understanding how the machine will work enough to calculate the lowest necessary specifications and build it.
3) Just as programs are typically written in high-level languages like Python, not in binary, machines are usually made from pre-made components that have specific functions.
And when we say “function” we mean like in pre-calculus as something that has inputs that produce predictable outputs. It may help to remember that functions are sometimes called “transformations”. The purpose of a component is to transform an input into the desired output. For example, a bike pump transforms your pushing motion into compressed air. If what you need is compressed air, and you have plenty of calories lying around in your body, this is a good trade. So think of a machine as a series of these “trades” chained together to get the final action at the end. A good machine gets from what you have (e.g. a wall outlet in your house that can provide 120V AC electricity) to the thing you want (e.g. hot food) as efficiently as possible.
The reason Rube Goldberg machines are funny is they are very inefficient, complex ways of accomplishing relatively simple tasks. They are also very educational for describing how these chains of components and chains of causality work, so I’ll use one as an example of how components are chained together to make a machine.
A) Transforms sitting force into compressed air.
B) Transforms compressed air into motion.
C) Transforms motion into contact.
D) Transforms contact into application of heat.
E) Transforms heat into loud noise.
F) Transforms loud noise into kinetic energy.
G) Transforms kinetic energy into signal to the camera.
As you can see, this is a lot like a syllogism of the form:
If A then B
If B then C
If C then D
If F then G.
If A then G.
When A is supplied with the correct input, and the machine is “functioning” (do you see now that the words we use for these things were not arbitrary choices?), then the predicted output of G is produced. Each step in this syllogism is a very simple function. Thus, it could be rewritten as:
A(x) = y or
A(input) = output
B(input) = output
G(input) = output
Like in math, a “system” of functions is a set or collection of functions that you deal with all together at once. That’s why we refer to things like mechanical “systems” or electro-hydraulic “systems”. All we’re really saying is there’s a chain of components performing mechanical functions (transforming mechanical things into other mechanical things) or a chain of components transforming electrical and hydraulic things into other electrical and hydraulic things.
Understanding how machines work is mostly a matter of understanding how the components work. Once you have the vocabulary (where the words are different kinds of components), any particular machine can then be understood as a sequence of logical statements of how the input of A eventually transforms into the output of G. Anything else about the machine is purely a matter of enabling that function to happen.
An example of enabling components is transmission systems. Transmission is usually just very simple components that take the output of A to the input of B because, due to space considerations, they couldn’t just be right next to each other. For example, one of those power drills that plugs into the wall needs a cable because if it just had a plug on it, you wouldn’t be able to move it around. The wire inside the cable is just a very simple component that transforms “electrical power over here” into “electrical power over there”. A garden hose turns “pressurized water over here” into “pressurized water over there”. The bike pump example from before has a pneumatic hose on it so you don’t have to plug the cylinder right into the bike plug.
Some transmission systems, like the one in the underbody of a car, can actually have pretty complicated working principles. But most of them are extremely simple. And the *function* of them is always simple: to transfer the output created by the component “over here” into the input needed “over there”. In the Rube Goldberg example, there are several examples of transmission components.
Another example of a consideration is how inputs and outputs of components will be connected. This is where the idea of conventions and standardization come in really handy. Probably the most important reason is because it prevents operators who don’t understand the machine from doing anything really stupid or dangerous by making it inconvenient. Even the dumbest person knows you don’t plug a USB cable directly into a wall socket. Why not? It’s just different kinds of wires, right? Well, it’s possible you could cut off the plug, strip the cable down to the bare wires, and then plug those into the wall. But because connections are conventional and standardized, we know this is probably dangerous, and whatever is in the wall socket is probably not the right input for whatever the USB is connected to. So even though, you might theorize “it’s just electricity”, it’s probably not the “right kind” of electricity. And you’d be right! If the person who made the machine wanted to supply it with 120V of AC current, he probably would have put a three-prong wall socket plug on it.
This is why European wall sockets are shaped differently from American wall sockets. If you tried to save money on the plug converter and stripped the cable of your laptop to plug it into the 110V AC wall plug, you’d be doing a very foolish thing. I don’t know what would happen exactly, but common sense says when you use a machine in a way that wasn’t intended you’re taking a huge, unnecessary personal risk unless you understand the machine as well as the original designer or better.
The other reason for standardizing connectors is because it makes it WAY more convenient to connect transmission components to other things. For example, you may have noticed it’s a lot easier to plug things into a wall socket than it is to hotwire them directly into the transformer outside your house. Even when you understand what you’re doing, and you know that your table saw takes 120V AC as an input, it’s still almost always a good idea to use a cable with a standard manufactured plug rather than plugging bare wires into the socket. Even though there’s no logical difference between the two (both will allow the machine to function), the conventional plug is better. Planning which connectors to use between components and transmission media saves a lot of trouble when it comes to buying things and then putting them together. Not planning this in detail is also, in my experience, the biggest and most predictable cause of headaches and negative consequences from last-minute hack jobs. Getting to the point of finally building the damn machine you’ve been planning for months and finding you have the wrong plug on a cable or the wrong screw thread on a hydraulic fitting is, apparently, a problem humanity has no intention of solving. (Until now!)
I’ll wait to describe the purpose of what I’d consider the last elements- structure, mounting, and fasteners- when I get around to the problems involved in spatial considerations. For now, let’s return to the figure I drew of a typical component and list everything we need to know about a component to understand it.
As mentioned before, a component minimally has a function, inputs, and outputs. If it works, then we don’t care how it works other than, perhaps, scientific curiosity. But as you may have guessed by now, components are often made with *gasp* other little components! That’s right, a component may be, itself, a little machine. I’ll give you a moment to compose yourself from the excitement of that revelation…
I always knew it would always come back to those goddamn turtles
Here, I’ll ask you to brush up on the idea of abstraction: https://aeolipera.wordpress.com/2017/04/01/abstraction/.
But some of the functions inside of a manufactured component may be performed by features that the manufacturer created from scratch in the manufacturing process. Maybe instead of little wires and tubes they drilled holes to allow fluids to pass between components, or trails of melted gold were drizzled across PCB boards to conduct small currents to get from A to B.
One of the coolest things to realize about manufactured things is that every feature, every detail, every bump, cranny, hole, flare, and rounded edge, was engineered for an explicit, specific purpose. Otherwise, they wouldn’t have gone to such outrageous expense to build a whole line of giant machines just to put little bumps, crannies, holes, flares, and rounded edges on things. Or at least I think this is a cool revelation. It completely changed the way I look at everyday manufactured objects. I see contours and think “What is that for? Why did they think that was important enough to pay for a machine to do that? What did it cost to manufacture it that way instead of a simpler way?”
Most of the machines that aren’t consumer goods are designed to make these little features as part of a manufacturing process. I’ll return to those processes to tie this all together someday, hopefully. Even the aesthetic features of consumer products, like the infamous rounded corners on iPhones, have to be machined that way at great expense. And even this is, arguably, a function. The input is the aesthetic, and the output is the feeling it induces in the consumer.
Anyway, returning to what’s going on inside of components, the knowledge that changes it from a black box you don’t care to understand to a little machine that you do understand is called the “working principle”. This is the part where physics becomes engineering. Before we can have transformations from mechanical things to thermodynamic things, or electrical things to optical things, and so on, somebody had to make a discovery in physics. An example from the lightbulb is, somebody had to realize that when current passes through a certain kind of wire, the resistance produces heat, which causes the wire to produce light. That process is the working principle of a lightbulb. At minimum, all you need to have something called a lightbulb is something that can use the input to operate the working principle of electrical resistance -> heat -> light. The rest- the glass bulb, the standardized screw-threaded plug on the bottom, the gas in the bulb- is just enabling the working principle to work better.
If some random, bizarre material were discovered to have the same physical property of electrical resistance as tungsten- let’s imagine a lab passed current through clay that had been soaked in worcestershire sauce and got the same number of Ohms as tungsten- then you could plug a lump of that thing into your wall and have a very odd looking “lightbulb”. But if you have a lightbulb that uses a different working principle, maybe by converting mechanical energy into heat, then you wouldn’t actually have a lightbulb at all even if you made it to look like one. It has a different working principle, so it’s a different kind of thing. Often the physical principle will be called an “X effect” like Bernoulli effect, Venturi effect, photoelectric effect, etc.
Components don’t exist in a perfect world where they just have functional inputs and outputs, they also have incidental inputs and outputs that are not intended and often negative. One of these is inputs from the environment. These can break the component’s function or subtly modify the intended, functional input and operation of the working principle in ways that weren’t designed and we probably wouldn’t like. For example, computers operated in space have to deal with high levels of radiation that can damage components and randomly flip bits back and forth. Unusual temperatures often effect physical changes in the component’s working principle or the transmission media. Even driving through a tunnel while talking on your phone is an example of this, because the transmission medium (EM waves) is reflected and transformed in ways the telecom system designer didn’t intend. These often require the design to include additional “enabling components” that don’t serve the machine’s function directly but they do ameliorate or prevent these unintended inputs. For example, space computers may have radiation-shielding enclosures, drinking water systems may have water filtration components, and so on.
There are also unintended outputs from each component referred to as “externalities”. Since these are often unwanted, we often call them “negative externalities”. Examples could be heat produced by a motor, pollution produced by a car, or even cow dung. Sometimes these are unwanted but also unavoidable, like excess heat that must be dispelled through a heat sink and cooling system, and sometimes these are avoidable but accidental (like the stuff getting into Flint’s drinking water from old pipes). Actually, in retrospect I think that may have been from some experimental treatment they added to the water. That would also be an example of a negative externality, except farther upstream in the process. Very rarely, an ingenius designer can use externalities as part of a larger system design. Maybe excess engine heat is used to heat the car’s interior or windows during the winter, or pollution is used to poison one’s neighbors to eliminate evolutionary competitors, or cow dung is sold as fertilizer.
In any case, understanding the environment a component is operating under and the planned and unplanned externalities it produces under normal and abnormal operation are very important for troubleshooting malfunctions from symptoms. Sometimes externalities affect the operating environment in a feedback loop. These special cases are called “considerations”. The easiest example is heat. Many machine components produce heat during operation which, obviously, change the temperature the component is operating in!
But depending on the abstraction level you’re looking at, there can be a lot of indirect examples. A big, plant-wide hydraulic system with a leak may cause one machine component’s operating environment to be “soaked in hydraulic fluid”. Since a machine is just a component in a bigger machine called a “production line”, and a production line is just a component in an even bigger machine called a “factory”, the number of considerations can be as large as your imagination. Fortunately, as with the original descriptions of function, it’s only necessary to be as comprehensive as is practically necessary to produce the understanding which, in turn, produces a properly functioning machine.