DesignWIKI

Fil Salustri's Design Site

Site Tools


design:detailing_a_design

Detailing a Design

Detailed design is where you start ascribing size, shape, and material to parts and assemblies.

Overview

At some point in a design process, the designers will actually have to specify the size, shape, material, tolerances, manufacturing information, and other detailed information that will allow their design to be realized. This stage of design is called detailed design.

Some engineers have taken the position that detailed design requires significant knowledge of highly technical and specialized topics; this would naturally make it quite difficult for easly-year students to carry out. However, I do not agree with this perspective. I believe that it is relatively easy to carry out a significant amount of detailed detail. Indeed, the Pareto Principle suggests that getting to within 80% of the “right” detailed design takes only about 20% of the effort. This means that we can expect students to carry out pretty good detailed design work.

It must be noted, of course, that the last 20% of the detailed design work – which takes 80% of the effort – is usually about optimizing a design to be as efficient, effective, and affordable as possible. It is in the last 20% of the detailed design that contracts are won or lost, and so it is very important. However, one must also remember that one can only get to the last 20% of the design once one has done the first 80%.

So, (a) students can do detailed design, and (b) they should do detailed design in order to understand just how important it is to product development.

First, let us summarize the design process that has been described so far.

  1. Design problem clarification and analysis. We studied the problem, brainstormed various ideas, and from those, established certain key descriptions of the design problem. This include all the requirements to properly describe the problem. At this point, no commitment is made to what the product will look like, but we do have a good understanding of what it is supposed to do.
  2. Systems design. In this stage, we interconnected functions into groups, thus defining systems in our product, and specified the kinds of interfaces needed to make the systems interact properly. At the end of this stage, we could envision the various sub-assemblies needed to make our product.
  3. Conceptual design and evaluation. In this stage, general concepts for the product as a whole were developed, refined, and evaluated. The evaluation is done with respect to criteria derived from the problem specification. At the end of this stage, a single general design idea is identified as the “lead solution” for the problem.

Now the concept has to be fleshed out – detailed – to the point where a manufacturer could actually make the product. That's what detailed design is.

From systems to parts

In “real-life” complex products (cars, boats, planes, etc.) each system can be treated like a separate design problem, and we could start a new design process for each of them. For example, in designing a car, one might start with designing the overall systems of a car, but then treat the design of the engine as a separate problem, and apply all the same methods to it. We can do this because of the recursion built into our systems engineering approach. The system interfaces between the engine and the rest of the car are of the same type as the interfaces between the car and the rest of the world. That is, a PRS describes the system inputs and outputs for the whole product.

In the design of automobiles, one can expect three or four levels of systems; in an airplane there can be six or seven.

However, in an academic setting where there is so little time to do this kind of recursive design, we must limit ourselves to only two levels of systems. The projects have been chosen specifically to encourage design teams to have only one set of subsystems in their design below the product itself – this is not a requirement, of course, because there is no single “right” answer in design.

One generally knows that there are no subsystems because one can easily envision a single component, be it a bolt, a lug, a motor, a hydraulic cylinder, and so on, to fill in the role of a systems component.

Note that the boundary of the system on the outside is the boundary between your product as a whole and the rest of the world. There can also be an “inner” boundary at the lowest level of your product's components. For example, if your company is designing a refrigerator, it will likely not also design the electric motor that the refrigerator will use. That is, your company designs refrigerators, not motors; you will likely just “spec” an appropriate motor that you found in a catalog of some sort. This means that you are not responsible for the design of the motor, only for selecting the right one. If you are not responsible for the motor, then the boundary between your product system ends at the motor.

Here is another example. Consider the design of a stapler for home use. We might easily envision the following systems:

  • Structural system: These are the components that carry all the loads that occur during the use of the stapler.
  • Stapling system: These are the components that actually do the stapling.
  • User interface: These are the components that allow a user to use the stapler safely and efficiently.
  • Storage system: These are the components that hold and deliver the staples to the stapling system.

The most important point here is this: the systems are distinct and separate, but the physical parts that implement the systems may be common to many systems. For example, in the typical stapler, the guide that carries the staples internally is a single piece of metal, but it is a part of the structural system because it carries many of the loads in the stapler, as well as part of the storage system because it holds the staples.

Yet another example is the typical automobile engine block. It is obviously part of the structural system of the engine – it holds the crank and many other parts in place. It is also part of the combustion system because it must have thermal properties that promote good combustion of fuel and air. It is also part of the lubrication system since there are typically passages (holes) in the engine block that let lubricants circulate from the top of the engine downwards.

Detailed design is about designing the actual parts of the product. So detailed design starts here, by looking at the PAS and concept design and starting to choose physical elements that can implement the functions of each system. Since each system is defined in terms of requirements of the whole product, each part ends up providing some of the functions needed to make the product work. So in deciding what kind of parts to use, you will often also find yourself looking back to your PRS.

Let's look at the stapler again. Using the principle of a class 2 lever makes sense because it provides the best trade-off between mechanical advantage (the increased force provided by the lever) and overall size of the stapler. This is why most manual staplers all look the same. Obviously, the lever itself must be a structural member (i.e. an element of the structural system). However, one also needs to store the staples. If you are designing a stapler, you will likely want it to work with staples as supplied in the most common form, which is in a long band of staples weakly bonded together. It doesn't take much to realise that as the lever will be strong and long and straight, it can also be used to carry the staples. Thus one can design the central structure of the stapler to do two things at once – i.e. to be part of two systems.

Some details about detailed design

Detailed design must provide the manufacturer with enough information to make all the parts needed for the product and to assemble the parts properly.

An important implication of this is that you may find an existing part in a catalog that is just right for your purposes (fasteners like bolts and screws are obvious examples). If a part can be ordered or purchased from an existing source, then there is no need to draw it. This is an important rule: do not design something if you can find something already designed. Let's say your product requires a gear. It makes economic sense to use an existing gear. In that case, you would find the appropriate gear, and specify that it be used in your design – you do not need to draw such a part separately.

The Thomas Register is an excellent resource for this; it contains product information from over 170,000 different companies. You can use it for free. You may register to use the advanced search features for free also, although this means you will occasionally receive an email message from the company.

Another good online catalog portal is TraceParts Online. There are others as well.

However, even if you find parts online, or in a hardcopy catalogue, you must still indicate how the part is to be assembled into the product. So all parts, even those for which detailed drawings are not provided, must be shown in the “assembly drawing” of your product. All parts, even the so-called off-the-shelf ones, must also be listed in your design's Bill of Materials.

Part design involves:

size & shape
What will the part look like? Should it be hollow?
The most significant trade-offs here will involve shape complexity vs weight: you can often lower a part's weight by increasing the complexity of a part's shape. But weight and complexity both affect manufacturing, sustainability, reliability, cost, etc.
materials
What will the part be made of? Material selection impacts weight, cost, manufacturability, reliability, sustainability, etc. Clearly, material selection is also very tightly coupled to size and shape.
connectivity
How will different parts connect to one another? Those physical connections are interfaces across which mass, energy, and information can flow. If the interfaces are mismatched, then reliability, robustness, safety, and many other characteristics will suffer.
operation
How will the part provide its intended function? How complex must its inputs and outputs be?

The features noted above will all impact many characteristics of the part:

manufacturability
Some parts are just easier (and therefore cheaper and more reliable) to make.
usability
Will a human be able to hold/grip/push/pull/lift/etc the part? (Not every part will be “used” by a human, but you might be surprised by how many parts actually are manipulated by a human at some point in their lifecycle.)
sustainability
What is the environmental impact of the part? Does an increased impact of a part's manufacture lower the environmental impact of its use? Can a lower environmental impact of one part help offset the higher environmental impact of another part?
maintenance
How easily can the part be maintained? Can it be removed/replaced quickly and with few (if any) special tools? How often should preventative maintenance be carried out? What is the cost of maintenance, and how does that compare to making a higher quality part that might need less maintenance?
end-of-life
What happens to the part when it's no longer needed? Can it be recycled or reused or remanufactured? Is it biodegradable without being a biohazard? How long will it take the part to degrade naturally?

Since a product's parts may be elements of more than one system, the impact of changing the design of one part can have all kinds of different impacts on other parts in other systems of the product. This means that you must pay careful attention to these interactions between parts. Here again, the PAS, and PRS documents can be very useful in helping a design team track all the possible variations that can arise just by changing one part. Never carry out detailed design without reference to the chosen concept and system. That is, know exactly what the part/component is supposed to do before designing it.

The act of defining a component of a system puts constraints on the system itself. In other words, designing parts adds new constraints to the design. Think about the restrictions each part places on the system, and make sure the restrictions won't adversely affect the overall design.

The simplest and cheapest component may not always be the most economic for the design overall. Take the time to consider the effect of an expensive part on the overall design; you may be surprised to find the overall product will be cheaper (in the long run and over the whole life of the product) if it has a few expensive parts in it than if it has only cheap parts in it.

Terms like “simple” and “cheap” can only have meaning with respect to a context. Your context is the design problem that you have set for yourself. For a team to develop a good design, all team members must agree on the context – on the design problem – so that everyone works with compatible definitions of important concepts like simplicity and affordability.

Generally, a reduction of component variety leads to shorter lead times, lower cost, and improved reliability.

Do not forget about manufacturability. While you may not be able to define the cost of manufacture precisely, you can still make some generalizations. For example, steel is heavier than aluminum but easier to work with and thus cheaper to use in manufacturing. How important is weight with respect to cost?

Make sure you've thought through how the product will be assembled. It is very easy to design assemblies that simply cannot be put together because of the order in which the parts must be assembled. Don't forget about installing fasteners – have you left enough room to insert nuts and bolts and the wrenches needed to tighten them? If you need to use a three inch long bolt to join two parts, is there room in the assembly for someone to slide the bolt into position and hold it with some kind of tool while the nut is being installed?

Safety factors (SFs) are also important. A A safety factor is a multiplier that increases the strength of a part to ensure it will be safe under the broadest feasible set of circumstances; for instance, a SF of 1.25 means that a part is 1.25 times stronger than it needs to be. While you may not be able to calculate the best SFs, you can still make a reasonable assumption. The most important thing to remember is that SFs will make your parts bigger and heavier, which will in turn change other aspects of your design, such as manufacturability, maintenance, and cost.

Finally, remember the KISS principle – simplicity is always good.

A simple method to help map function to form

One of the great challenges of design is the conversion from function (what the product is supposed to do) to form (what the product has to look like). There is no single perfect method for this. Creativity and innovation figures prominently here. However, here are some simple guidelines that should help you map function to form.

  1. For each function to be provided by a system, list several different concepts that describe a potential physical component. For example, if the required function is to convert linear motion into circular motion, two possible concepts are a slider-crank mechanism, or a linear induction motor. These are very different concepts.
  2. Consider the product characteristics and constraints with respect to each of your concepts. Evaluate which of the concepts best satisfies the PCs and constraints.
  3. Rate the concepts by the results of the above evaluation and choose the best concept.

Notice that this is essentially the same method as was used in concept design. You could use the decision matrix method here, just as was done for the concept design stage, but it is not required for your project.

Some notes on project management during detailed design:

  • If one team member is in charge of each system, then that same team member should be in charge of doing the detailed design of that system.
  • Since systems interact, it's important to get the team members responsible the detailed design of pertinent co-systems involved.
    • Example: The structural system and power system of a bicycle interact. It makes sense, then, for the person detailing the structural design to seek the advice of the person detailing the power system before making any significant decisions about bicycle structure, especially insofar as the interfaces between the structural system and power system are concerned.

Load lines: a simple qualitative stress analysis

Even though you may not yet know enough math and physics to conduct a proper analysis of the structure of your product, you can make some very useful qualitative analyses. Using the concept of load lines is something that is useful and that you can do without knowledge all the math and just a little physics.

The idea is that loads of any kind (electrical, mechanical, thermal, etc) “flow” much like water does: in a straight line unless it's deflected by something. Imagine water flowing through a pipe that suddenly gets wider. The water will fill the pipe both before and after the change in width; it will slow down where the pipe is wider, and speed up where the pipe is narrower. Also, think about water flowing over a step; it will keep at a fairly constant level before the step, but it will not turn a sudden right-angle at the step; rather, it will flow around the step smoothly.

The same thing happens to forces applied to structures. Say you have a square steel plate with sharp corners, and you apply an external load to two adjacent sides of the plate, in the plane of the plate. The load will need to “move” through the plate. It will flow smoothly, even around the corners.

Fig. 1: Examples of load lines.

Some examples of load lines are given in Figure 1. Look at the top left example. Imagine that it shows the top view of a channel that is horizontal; imagine that the arrows at the bottom of the channel indicate the direction of flow of water entering the channel. Imagine the water flowing slowly through the channel and around the corner. Now, imagine you dropped a light plastic bead into the flow, and watched it as it is carried by the water current. It would trace out lines similar to the dashed lines in the figure. Notice that the lines:

  • are smooth – the bead does not suddenly turn a sharp corner
  • never cross one another
  • enter and exit the channel parallel to the flow arrows in and out of the channel
  • never leave except via another arrow
  • form the shortest reasonable path from entrance to exit without violating the other rules.

Now think of the same geometry – not as a channel, but rather as a plate or block of some sort. The arrows indicate where the applied loads and reaction loads are applied. We assume the object is at equilibrium (it is not accelerating, rotating, etc.). The very same lines that marked the movement of the beads carried by the water also indicate how the forces moves through the analogous solid part.

The more bunched up the lines are, the more force you have in any given region of the part. In the top, right part of Figure 1, the lines bunch up the most on the inside of the corner. That means this is where the forces will be most concentrated, which in turn means that the inside corner is the region where the part is most likely to break. Similarly, the lines of flow at the outer corner are the most spread out. This means that this is where the forces are lowest, and where there is the least chance of the part breaking.

What does this mean with respect to your design? It can be used to “tune” the design of the part's shape to match the forces that it will have to endure. Since the force lines bunch up at the sharp inside corner, it would help the design to remove the corner – that is, to replace it with a fillet. The bigger the fillet, the more well distributed the forces will be, and the less likely the part's failure will be. Also, don't forget that the part has a certain thickness (measured here out of the plane of the drawing). In order to make the part support forces in the inside corner, you might also consider making it thicker. This provides more material to carry those forces without changing the planar geometry of the part.

Furthermore, since the forces are very low at the outer corner, there is little need of all that material. You can improve this part's design by putting a chamfer on that corner. The chamfer effectively removes material since it isn't needed. This makes the part lighter and cheaper due to the lesser amount of material needed.

This technique doesn't only work with physical loads; it also works with thermal loads. Think of the shapes in Figure 1 as representing parts through which heat flows. The arrows represent sources and sinks for the heat. The other sides are considered perfectly insulated. (This is a big assumption, of course, but remember that we are only approximating the general behaviour of parts here.) In this case, heat will flow through the part following the dashed lines. The lines bunch up in areas of higher temperature and are spread out in areas of lower temperature. By thinking about how the heat flows, you may decide where most of the insulation needs to be placed, or which materials to use to ensure the heat is transferred well through the entire part.

design/detailing_a_design.txt · Last modified: 2020.03.12 13:30 (external edit)